id
large_stringlengths 9
16
| submitter
large_stringlengths 3
64
⌀ | authors
large_stringlengths 5
20.7k
| title
large_stringlengths 5
245
| comments
large_stringlengths 1
482
⌀ | journal-ref
large_stringlengths 4
382
⌀ | doi
large_stringlengths 9
151
⌀ | report-no
large_stringlengths 2
509
⌀ | categories
large_stringlengths 5
78
| license
large_stringclasses 9
values | abstract
large_stringlengths 83
3.55k
| versions
large listlengths 1
21
| update_date
timestamp[ms]date 2007-05-23 00:00:00
2025-04-25 00:00:00
| authors_parsed
large listlengths 1
1.35k
| prompt
large_stringlengths 166
3.64k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2303.01192 | Joon-Hyuk Chang | Ye-Rin Jeoung, Joon-Young Yang, Jeong-Hwan Choi and Joon-Hyuk Chang | Improving Transformer-based End-to-End Speaker Diarization by Assigning
Auxiliary Losses to Attention Heads | 5 pages, ICASSP2023 (Accepted) | null | null | null | eess.AS cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transformer-based end-to-end neural speaker diarization (EEND) models utilize
the multi-head self-attention (SA) mechanism to enable accurate speaker label
prediction in overlapped speech regions. In this study, to enhance the training
effectiveness of SA-EEND models, we propose the use of auxiliary losses for the
SA heads of the transformer layers. Specifically, we assume that the attention
weight matrices of an SA layer are redundant if their patterns are similar to
those of the identity matrix. We then explicitly constrain such matrices to
exhibit specific speaker activity patterns relevant to voice activity detection
or overlapped speech detection tasks. Consequently, we expect the proposed
auxiliary losses to guide the transformer layers to exhibit more diverse
patterns in the attention weights, thereby reducing the assumed redundancies in
the SA heads. The effectiveness of the proposed method is demonstrated using
the simulated and CALLHOME datasets for two-speaker diarization tasks, reducing
the diarization error rate of the conventional SA-EEND model by 32.58% and
17.11%, respectively.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 12:15:36 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Jeoung",
"Ye-Rin",
""
],
[
"Yang",
"Joon-Young",
""
],
[
"Choi",
"Jeong-Hwan",
""
],
[
"Chang",
"Joon-Hyuk",
""
]
] | TITLE: Improving Transformer-based End-to-End Speaker Diarization by Assigning
Auxiliary Losses to Attention Heads
ABSTRACT: Transformer-based end-to-end neural speaker diarization (EEND) models utilize
the multi-head self-attention (SA) mechanism to enable accurate speaker label
prediction in overlapped speech regions. In this study, to enhance the training
effectiveness of SA-EEND models, we propose the use of auxiliary losses for the
SA heads of the transformer layers. Specifically, we assume that the attention
weight matrices of an SA layer are redundant if their patterns are similar to
those of the identity matrix. We then explicitly constrain such matrices to
exhibit specific speaker activity patterns relevant to voice activity detection
or overlapped speech detection tasks. Consequently, we expect the proposed
auxiliary losses to guide the transformer layers to exhibit more diverse
patterns in the attention weights, thereby reducing the assumed redundancies in
the SA heads. The effectiveness of the proposed method is demonstrated using
the simulated and CALLHOME datasets for two-speaker diarization tasks, reducing
the diarization error rate of the conventional SA-EEND model by 32.58% and
17.11%, respectively.
|
2303.01196 | Houssem Eddine Boulahbal | Houssem Boulahbal, Adrian Voicila, Andrew Comport | STDepthFormer: Predicting Spatio-temporal Depth from Video with a
Self-supervised Transformer Model | Submitted to IROS 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, a self-supervised model that simultaneously predicts a
sequence of future frames from video-input with a novel spatial-temporal
attention (ST) network is proposed. The ST transformer network allows
constraining both temporal consistency across future frames whilst constraining
consistency across spatial objects in the image at different scales. This was
not the case in prior works for depth prediction, which focused on predicting a
single frame as output. The proposed model leverages prior scene knowledge such
as object shape and texture similar to single-image depth inference methods,
whilst also constraining the motion and geometry from a sequence of input
images. Apart from the transformer architecture, one of the main contributions
with respect to prior works lies in the objective function that enforces
spatio-temporal consistency across a sequence of output frames rather than a
single output frame. As will be shown, this results in more accurate and robust
depth sequence forecasting. The model achieves highly accurate depth
forecasting results that outperform existing baselines on the KITTI benchmark.
Extensive ablation studies were performed to assess the effectiveness of the
proposed techniques. One remarkable result of the proposed model is that it is
implicitly capable of forecasting the motion of objects in the scene, rather
than requiring complex models involving multi-object detection, segmentation
and tracking.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 12:22:51 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Boulahbal",
"Houssem",
""
],
[
"Voicila",
"Adrian",
""
],
[
"Comport",
"Andrew",
""
]
] | TITLE: STDepthFormer: Predicting Spatio-temporal Depth from Video with a
Self-supervised Transformer Model
ABSTRACT: In this paper, a self-supervised model that simultaneously predicts a
sequence of future frames from video-input with a novel spatial-temporal
attention (ST) network is proposed. The ST transformer network allows
constraining both temporal consistency across future frames whilst constraining
consistency across spatial objects in the image at different scales. This was
not the case in prior works for depth prediction, which focused on predicting a
single frame as output. The proposed model leverages prior scene knowledge such
as object shape and texture similar to single-image depth inference methods,
whilst also constraining the motion and geometry from a sequence of input
images. Apart from the transformer architecture, one of the main contributions
with respect to prior works lies in the objective function that enforces
spatio-temporal consistency across a sequence of output frames rather than a
single output frame. As will be shown, this results in more accurate and robust
depth sequence forecasting. The model achieves highly accurate depth
forecasting results that outperform existing baselines on the KITTI benchmark.
Extensive ablation studies were performed to assess the effectiveness of the
proposed techniques. One remarkable result of the proposed model is that it is
implicitly capable of forecasting the motion of objects in the scene, rather
than requiring complex models involving multi-object detection, segmentation
and tracking.
|
2303.01197 | Kashif Ahmad | Muhammad Tayyab Zamir, Muhammad Asif Ayub, Jebran Khan, Muhammad Jawad
Ikram, Nasir Ahmad, Kashif Ahmad | Document Provenance and Authentication through Authorship Classification | 7 pages; 3 tables; 1 figure | IEEE ICAISC 2023 | null | null | cs.CL cs.DL | http://creativecommons.org/licenses/by/4.0/ | Style analysis, which is relatively a less explored topic, enables several
interesting applications. For instance, it allows authors to adjust their
writing style to produce a more coherent document in collaboration. Similarly,
style analysis can also be used for document provenance and authentication as a
primary step. In this paper, we propose an ensemble-based text-processing
framework for the classification of single and multi-authored documents, which
is one of the key tasks in style analysis. The proposed framework incorporates
several state-of-the-art text classification algorithms including classical
Machine Learning (ML) algorithms, transformers, and deep learning algorithms
both individually and in merit-based late fusion. For the merit-based late
fusion, we employed several weight optimization and selection methods to assign
merit-based weights to the individual text classification algorithms. We also
analyze the impact of the characters on the task that are usually excluded in
NLP applications during pre-processing by conducting experiments on both clean
and un-clean data. The proposed framework is evaluated on a large-scale
benchmark dataset, significantly improving performance over the existing
solutions.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 12:26:03 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Zamir",
"Muhammad Tayyab",
""
],
[
"Ayub",
"Muhammad Asif",
""
],
[
"Khan",
"Jebran",
""
],
[
"Ikram",
"Muhammad Jawad",
""
],
[
"Ahmad",
"Nasir",
""
],
[
"Ahmad",
"Kashif",
""
]
] | TITLE: Document Provenance and Authentication through Authorship Classification
ABSTRACT: Style analysis, which is relatively a less explored topic, enables several
interesting applications. For instance, it allows authors to adjust their
writing style to produce a more coherent document in collaboration. Similarly,
style analysis can also be used for document provenance and authentication as a
primary step. In this paper, we propose an ensemble-based text-processing
framework for the classification of single and multi-authored documents, which
is one of the key tasks in style analysis. The proposed framework incorporates
several state-of-the-art text classification algorithms including classical
Machine Learning (ML) algorithms, transformers, and deep learning algorithms
both individually and in merit-based late fusion. For the merit-based late
fusion, we employed several weight optimization and selection methods to assign
merit-based weights to the individual text classification algorithms. We also
analyze the impact of the characters on the task that are usually excluded in
NLP applications during pre-processing by conducting experiments on both clean
and un-clean data. The proposed framework is evaluated on a large-scale
benchmark dataset, significantly improving performance over the existing
solutions.
|
2303.01201 | Zhen Cheng | Zhen Cheng, Fei Zhu, Xu-Yao Zhang, Cheng-Lin Liu | Average of Pruning: Improving Performance and Stability of
Out-of-Distribution Detection | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Detecting Out-of-distribution (OOD) inputs have been a critical issue for
neural networks in the open world. However, the unstable behavior of OOD
detection along the optimization trajectory during training has not been
explored clearly. In this paper, we first find the performance of OOD detection
suffers from overfitting and instability during training: 1) the performance
could decrease when the training error is near zero, and 2) the performance
would vary sharply in the final stage of training. Based on our findings, we
propose Average of Pruning (AoP), consisting of model averaging and pruning, to
mitigate the unstable behaviors. Specifically, model averaging can help achieve
a stable performance by smoothing the landscape, and pruning is certified to
eliminate the overfitting by eliminating redundant features. Comprehensive
experiments on various datasets and architectures are conducted to verify the
effectiveness of our method.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 12:34:38 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Cheng",
"Zhen",
""
],
[
"Zhu",
"Fei",
""
],
[
"Zhang",
"Xu-Yao",
""
],
[
"Liu",
"Cheng-Lin",
""
]
] | TITLE: Average of Pruning: Improving Performance and Stability of
Out-of-Distribution Detection
ABSTRACT: Detecting Out-of-distribution (OOD) inputs have been a critical issue for
neural networks in the open world. However, the unstable behavior of OOD
detection along the optimization trajectory during training has not been
explored clearly. In this paper, we first find the performance of OOD detection
suffers from overfitting and instability during training: 1) the performance
could decrease when the training error is near zero, and 2) the performance
would vary sharply in the final stage of training. Based on our findings, we
propose Average of Pruning (AoP), consisting of model averaging and pruning, to
mitigate the unstable behaviors. Specifically, model averaging can help achieve
a stable performance by smoothing the landscape, and pruning is certified to
eliminate the overfitting by eliminating redundant features. Comprehensive
experiments on various datasets and architectures are conducted to verify the
effectiveness of our method.
|
2303.01211 | Jun Xue | Jun Xue, Cunhang Fan, Jiangyan Yi, Chenglong Wang, Zhengqi Wen, Dan
Zhang, Zhao Lv | Learning From Yourself: A Self-Distillation Method for Fake Speech
Detection | Accepted by ICASSP 2023 | null | null | null | cs.SD cs.LG cs.MM eess.AS | http://creativecommons.org/licenses/by/4.0/ | In this paper, we propose a novel self-distillation method for fake speech
detection (FSD), which can significantly improve the performance of FSD without
increasing the model complexity. For FSD, some fine-grained information is very
important, such as spectrogram defects, mute segments, and so on, which are
often perceived by shallow networks. However, shallow networks have much noise,
which can not capture this very well. To address this problem, we propose using
the deepest network instruct shallow network for enhancing shallow networks.
Specifically, the networks of FSD are divided into several segments, the
deepest network being used as the teacher model, and all shallow networks
become multiple student models by adding classifiers. Meanwhile, the
distillation path between the deepest network feature and shallow network
features is used to reduce the feature difference. A series of experimental
results on the ASVspoof 2019 LA and PA datasets show the effectiveness of the
proposed method, with significant improvements compared to the baseline.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 12:52:22 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Xue",
"Jun",
""
],
[
"Fan",
"Cunhang",
""
],
[
"Yi",
"Jiangyan",
""
],
[
"Wang",
"Chenglong",
""
],
[
"Wen",
"Zhengqi",
""
],
[
"Zhang",
"Dan",
""
],
[
"Lv",
"Zhao",
""
]
] | TITLE: Learning From Yourself: A Self-Distillation Method for Fake Speech
Detection
ABSTRACT: In this paper, we propose a novel self-distillation method for fake speech
detection (FSD), which can significantly improve the performance of FSD without
increasing the model complexity. For FSD, some fine-grained information is very
important, such as spectrogram defects, mute segments, and so on, which are
often perceived by shallow networks. However, shallow networks have much noise,
which can not capture this very well. To address this problem, we propose using
the deepest network instruct shallow network for enhancing shallow networks.
Specifically, the networks of FSD are divided into several segments, the
deepest network being used as the teacher model, and all shallow networks
become multiple student models by adding classifiers. Meanwhile, the
distillation path between the deepest network feature and shallow network
features is used to reduce the feature difference. A series of experimental
results on the ASVspoof 2019 LA and PA datasets show the effectiveness of the
proposed method, with significant improvements compared to the baseline.
|
2303.01217 | Stefanos-Iordanis Papadopoulos | Stefanos-Iordanis Papadopoulos, Christos Koutlis, Symeon Papadopoulos,
Panagiotis C. Petrantonakis | Synthetic Misinformers: Generating and Combating Multimodal
Misinformation | null | null | null | null | cs.MM cs.AI cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | With the expansion of social media and the increasing dissemination of
multimedia content, the spread of misinformation has become a major concern.
This necessitates effective strategies for multimodal misinformation detection
(MMD) that detect whether the combination of an image and its accompanying text
could mislead or misinform. Due to the data-intensive nature of deep neural
networks and the labor-intensive process of manual annotation, researchers have
been exploring various methods for automatically generating synthetic
multimodal misinformation - which we refer to as Synthetic Misinformers - in
order to train MMD models. However, limited evaluation on real-world
misinformation and a lack of comparisons with other Synthetic Misinformers
makes difficult to assess progress in the field. To address this, we perform a
comparative study on existing and new Synthetic Misinformers that involves (1)
out-of-context (OOC) image-caption pairs, (2) cross-modal named entity
inconsistency (NEI) as well as (3) hybrid approaches and we evaluate them
against real-world misinformation; using the COSMOS benchmark. The comparative
study showed that our proposed CLIP-based Named Entity Swapping can lead to MMD
models that surpass other OOC and NEI Misinformers in terms of multimodal
accuracy and that hybrid approaches can lead to even higher detection accuracy.
Nevertheless, after alleviating information leakage from the COSMOS evaluation
protocol, low Sensitivity scores indicate that the task is significantly more
challenging than previous studies suggested. Finally, our findings showed that
NEI-based Synthetic Misinformers tend to suffer from a unimodal bias, where
text-only MMDs can outperform multimodal ones.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 12:59:01 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Papadopoulos",
"Stefanos-Iordanis",
""
],
[
"Koutlis",
"Christos",
""
],
[
"Papadopoulos",
"Symeon",
""
],
[
"Petrantonakis",
"Panagiotis C.",
""
]
] | TITLE: Synthetic Misinformers: Generating and Combating Multimodal
Misinformation
ABSTRACT: With the expansion of social media and the increasing dissemination of
multimedia content, the spread of misinformation has become a major concern.
This necessitates effective strategies for multimodal misinformation detection
(MMD) that detect whether the combination of an image and its accompanying text
could mislead or misinform. Due to the data-intensive nature of deep neural
networks and the labor-intensive process of manual annotation, researchers have
been exploring various methods for automatically generating synthetic
multimodal misinformation - which we refer to as Synthetic Misinformers - in
order to train MMD models. However, limited evaluation on real-world
misinformation and a lack of comparisons with other Synthetic Misinformers
makes difficult to assess progress in the field. To address this, we perform a
comparative study on existing and new Synthetic Misinformers that involves (1)
out-of-context (OOC) image-caption pairs, (2) cross-modal named entity
inconsistency (NEI) as well as (3) hybrid approaches and we evaluate them
against real-world misinformation; using the COSMOS benchmark. The comparative
study showed that our proposed CLIP-based Named Entity Swapping can lead to MMD
models that surpass other OOC and NEI Misinformers in terms of multimodal
accuracy and that hybrid approaches can lead to even higher detection accuracy.
Nevertheless, after alleviating information leakage from the COSMOS evaluation
protocol, low Sensitivity scores indicate that the task is significantly more
challenging than previous studies suggested. Finally, our findings showed that
NEI-based Synthetic Misinformers tend to suffer from a unimodal bias, where
text-only MMDs can outperform multimodal ones.
|
2303.01233 | Kaiyu Guo | Kaiyu Guo, Brian Lovell | Domain-aware Triplet loss in Domain Generalization | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Despite much progress being made in the field of object recognition with the
advances of deep learning, there are still several factors negatively affecting
the performance of deep learning models. Domain shift is one of these factors
and is caused by discrepancies in the distributions of the testing and training
data. In this paper, we focus on the problem of compact feature clustering in
domain generalization to help optimize the embedding space from multi-domain
data. We design a domainaware triplet loss for domain generalization to help
the model to not only cluster similar semantic features, but also to disperse
features arising from the domain. Unlike previous methods focusing on
distribution alignment, our algorithm is designed to disperse domain
information in the embedding space. The basic idea is motivated based on the
assumption that embedding features can be clustered based on domain
information, which is mathematically and empirically supported in this paper.
In addition, during our exploration of feature clustering in domain
generalization, we note that factors affecting the convergence of metric
learning loss in domain generalization are more important than the pre-defined
domains. To solve this issue, we utilize two methods to normalize the embedding
space, reducing the internal covariate shift of the embedding features. The
ablation study demonstrates the effectiveness of our algorithm. Moreover, the
experiments on the benchmark datasets, including PACS, VLCS and Office-Home,
show that our method outperforms related methods focusing on domain
discrepancy. In particular, our results on RegnetY-16 are significantly better
than state-of-the-art methods on the benchmark datasets. Our code will be
released at https://github.com/workerbcd/DCT
| [
{
"version": "v1",
"created": "Wed, 1 Mar 2023 14:02:01 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Guo",
"Kaiyu",
""
],
[
"Lovell",
"Brian",
""
]
] | TITLE: Domain-aware Triplet loss in Domain Generalization
ABSTRACT: Despite much progress being made in the field of object recognition with the
advances of deep learning, there are still several factors negatively affecting
the performance of deep learning models. Domain shift is one of these factors
and is caused by discrepancies in the distributions of the testing and training
data. In this paper, we focus on the problem of compact feature clustering in
domain generalization to help optimize the embedding space from multi-domain
data. We design a domainaware triplet loss for domain generalization to help
the model to not only cluster similar semantic features, but also to disperse
features arising from the domain. Unlike previous methods focusing on
distribution alignment, our algorithm is designed to disperse domain
information in the embedding space. The basic idea is motivated based on the
assumption that embedding features can be clustered based on domain
information, which is mathematically and empirically supported in this paper.
In addition, during our exploration of feature clustering in domain
generalization, we note that factors affecting the convergence of metric
learning loss in domain generalization are more important than the pre-defined
domains. To solve this issue, we utilize two methods to normalize the embedding
space, reducing the internal covariate shift of the embedding features. The
ablation study demonstrates the effectiveness of our algorithm. Moreover, the
experiments on the benchmark datasets, including PACS, VLCS and Office-Home,
show that our method outperforms related methods focusing on domain
discrepancy. In particular, our results on RegnetY-16 are significantly better
than state-of-the-art methods on the benchmark datasets. Our code will be
released at https://github.com/workerbcd/DCT
|
2303.01237 | Zhaoyang Huang | Xiaoyu Shi, Zhaoyang Huang, Dasong Li, Manyuan Zhang, Ka Chun Cheung,
Simon See, Hongwei Qin, Jifeng Dai, Hongsheng Li | FlowFormer++: Masked Cost Volume Autoencoding for Pretraining Optical
Flow Estimation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | FlowFormer introduces a transformer architecture into optical flow estimation
and achieves state-of-the-art performance. The core component of FlowFormer is
the transformer-based cost-volume encoder. Inspired by the recent success of
masked autoencoding (MAE) pretraining in unleashing transformers' capacity of
encoding visual representation, we propose Masked Cost Volume Autoencoding
(MCVA) to enhance FlowFormer by pretraining the cost-volume encoder with a
novel MAE scheme. Firstly, we introduce a block-sharing masking strategy to
prevent masked information leakage, as the cost maps of neighboring source
pixels are highly correlated. Secondly, we propose a novel pre-text
reconstruction task, which encourages the cost-volume encoder to aggregate
long-range information and ensures pretraining-finetuning consistency. We also
show how to modify the FlowFormer architecture to accommodate masks during
pretraining. Pretrained with MCVA, FlowFormer++ ranks 1st among published
methods on both Sintel and KITTI-2015 benchmarks. Specifically, FlowFormer++
achieves 1.07 and 1.94 average end-point error (AEPE) on the clean and final
pass of Sintel benchmark, leading to 7.76\% and 7.18\% error reductions from
FlowFormer. FlowFormer++ obtains 4.52 F1-all on the KITTI-2015 test set,
improving FlowFormer by 0.16.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 13:28:07 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Shi",
"Xiaoyu",
""
],
[
"Huang",
"Zhaoyang",
""
],
[
"Li",
"Dasong",
""
],
[
"Zhang",
"Manyuan",
""
],
[
"Cheung",
"Ka Chun",
""
],
[
"See",
"Simon",
""
],
[
"Qin",
"Hongwei",
""
],
[
"Dai",
"Jifeng",
""
],
[
"Li",
"Hongsheng",
""
]
] | TITLE: FlowFormer++: Masked Cost Volume Autoencoding for Pretraining Optical
Flow Estimation
ABSTRACT: FlowFormer introduces a transformer architecture into optical flow estimation
and achieves state-of-the-art performance. The core component of FlowFormer is
the transformer-based cost-volume encoder. Inspired by the recent success of
masked autoencoding (MAE) pretraining in unleashing transformers' capacity of
encoding visual representation, we propose Masked Cost Volume Autoencoding
(MCVA) to enhance FlowFormer by pretraining the cost-volume encoder with a
novel MAE scheme. Firstly, we introduce a block-sharing masking strategy to
prevent masked information leakage, as the cost maps of neighboring source
pixels are highly correlated. Secondly, we propose a novel pre-text
reconstruction task, which encourages the cost-volume encoder to aggregate
long-range information and ensures pretraining-finetuning consistency. We also
show how to modify the FlowFormer architecture to accommodate masks during
pretraining. Pretrained with MCVA, FlowFormer++ ranks 1st among published
methods on both Sintel and KITTI-2015 benchmarks. Specifically, FlowFormer++
achieves 1.07 and 1.94 average end-point error (AEPE) on the clean and final
pass of Sintel benchmark, leading to 7.76\% and 7.18\% error reductions from
FlowFormer. FlowFormer++ obtains 4.52 F1-all on the KITTI-2015 test set,
improving FlowFormer by 0.16.
|
2303.01249 | Zhijie Shen | Zhijie Shen, Wu Guo, Bin Gu | Language-Universal Adapter Learning with Knowledge Distillation for
End-to-End Multilingual Speech Recognition | null | null | null | null | cs.CL cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | In this paper, we propose a language-universal adapter learning framework
based on a pre-trained model for end-to-end multilingual automatic speech
recognition (ASR). For acoustic modeling, the wav2vec 2.0 pre-trained model is
fine-tuned by inserting language-specific and language-universal adapters. An
online knowledge distillation is then used to enable the language-universal
adapters to learn both language-specific and universal features. The linguistic
information confusion is also reduced by leveraging language identifiers
(LIDs). With LIDs we perform a position-wise modification on the multi-head
attention outputs. In the inference procedure, the language-specific adapters
are removed while the language-universal adapters are kept activated. The
proposed method improves the recognition accuracy and addresses the linear
increase of the number of adapters' parameters with the number of languages in
common multilingual ASR systems. Experiments on the BABEL dataset confirm the
effectiveness of the proposed framework. Compared to the conventional
multilingual model, a 3.3% absolute error rate reduction is achieved. The code
is available at: https://github.com/shen9712/UniversalAdapterLearning.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 14:43:49 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Shen",
"Zhijie",
""
],
[
"Guo",
"Wu",
""
],
[
"Gu",
"Bin",
""
]
] | TITLE: Language-Universal Adapter Learning with Knowledge Distillation for
End-to-End Multilingual Speech Recognition
ABSTRACT: In this paper, we propose a language-universal adapter learning framework
based on a pre-trained model for end-to-end multilingual automatic speech
recognition (ASR). For acoustic modeling, the wav2vec 2.0 pre-trained model is
fine-tuned by inserting language-specific and language-universal adapters. An
online knowledge distillation is then used to enable the language-universal
adapters to learn both language-specific and universal features. The linguistic
information confusion is also reduced by leveraging language identifiers
(LIDs). With LIDs we perform a position-wise modification on the multi-head
attention outputs. In the inference procedure, the language-specific adapters
are removed while the language-universal adapters are kept activated. The
proposed method improves the recognition accuracy and addresses the linear
increase of the number of adapters' parameters with the number of languages in
common multilingual ASR systems. Experiments on the BABEL dataset confirm the
effectiveness of the proposed framework. Compared to the conventional
multilingual model, a 3.3% absolute error rate reduction is achieved. The code
is available at: https://github.com/shen9712/UniversalAdapterLearning.
|
2303.01256 | Gautam Kamath | Xin Gu, Gautam Kamath, Zhiwei Steven Wu | Choosing Public Datasets for Private Machine Learning via Gradient
Subspace Distance | null | null | null | null | stat.ML cs.CR cs.CV cs.DS cs.LG | http://creativecommons.org/licenses/by/4.0/ | Differentially private stochastic gradient descent privatizes model training
by injecting noise into each iteration, where the noise magnitude increases
with the number of model parameters. Recent works suggest that we can reduce
the noise by leveraging public data for private machine learning, by projecting
gradients onto a subspace prescribed by the public data. However, given a
choice of public datasets, it is not a priori clear which one may be most
appropriate for the private task. We give an algorithm for selecting a public
dataset by measuring a low-dimensional subspace distance between gradients of
the public and private examples. We provide theoretical analysis demonstrating
that the excess risk scales with this subspace distance. This distance is easy
to compute and robust to modifications in the setting. Empirical evaluation
shows that trained model accuracy is monotone in this distance.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 13:36:28 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Gu",
"Xin",
""
],
[
"Kamath",
"Gautam",
""
],
[
"Wu",
"Zhiwei Steven",
""
]
] | TITLE: Choosing Public Datasets for Private Machine Learning via Gradient
Subspace Distance
ABSTRACT: Differentially private stochastic gradient descent privatizes model training
by injecting noise into each iteration, where the noise magnitude increases
with the number of model parameters. Recent works suggest that we can reduce
the noise by leveraging public data for private machine learning, by projecting
gradients onto a subspace prescribed by the public data. However, given a
choice of public datasets, it is not a priori clear which one may be most
appropriate for the private task. We give an algorithm for selecting a public
dataset by measuring a low-dimensional subspace distance between gradients of
the public and private examples. We provide theoretical analysis demonstrating
that the excess risk scales with this subspace distance. This distance is easy
to compute and robust to modifications in the setting. Empirical evaluation
shows that trained model accuracy is monotone in this distance.
|
2303.01263 | Enyan Dai | Enyan Dai, Minhua Lin, Xiang Zhang, and Suhang Wang | Unnoticeable Backdoor Attacks on Graph Neural Networks | null | null | 10.1145/3543507.3583392 | null | cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph Neural Networks (GNNs) have achieved promising results in various tasks
such as node classification and graph classification. Recent studies find that
GNNs are vulnerable to adversarial attacks. However, effective backdoor attacks
on graphs are still an open problem. In particular, backdoor attack poisons the
graph by attaching triggers and the target class label to a set of nodes in the
training graph. The backdoored GNNs trained on the poisoned graph will then be
misled to predict test nodes to target class once attached with triggers.
Though there are some initial efforts in graph backdoor attacks, our empirical
analysis shows that they may require a large attack budget for effective
backdoor attacks and the injected triggers can be easily detected and pruned.
Therefore, in this paper, we study a novel problem of unnoticeable graph
backdoor attacks with limited attack budget. To fully utilize the attack
budget, we propose to deliberately select the nodes to inject triggers and
target class labels in the poisoning phase. An adaptive trigger generator is
deployed to obtain effective triggers that are difficult to be noticed.
Extensive experiments on real-world datasets against various defense strategies
demonstrate the effectiveness of our proposed method in conducting effective
unnoticeable backdoor attacks.
| [
{
"version": "v1",
"created": "Sat, 11 Feb 2023 01:50:58 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Dai",
"Enyan",
""
],
[
"Lin",
"Minhua",
""
],
[
"Zhang",
"Xiang",
""
],
[
"Wang",
"Suhang",
""
]
] | TITLE: Unnoticeable Backdoor Attacks on Graph Neural Networks
ABSTRACT: Graph Neural Networks (GNNs) have achieved promising results in various tasks
such as node classification and graph classification. Recent studies find that
GNNs are vulnerable to adversarial attacks. However, effective backdoor attacks
on graphs are still an open problem. In particular, backdoor attack poisons the
graph by attaching triggers and the target class label to a set of nodes in the
training graph. The backdoored GNNs trained on the poisoned graph will then be
misled to predict test nodes to target class once attached with triggers.
Though there are some initial efforts in graph backdoor attacks, our empirical
analysis shows that they may require a large attack budget for effective
backdoor attacks and the injected triggers can be easily detected and pruned.
Therefore, in this paper, we study a novel problem of unnoticeable graph
backdoor attacks with limited attack budget. To fully utilize the attack
budget, we propose to deliberately select the nodes to inject triggers and
target class labels in the poisoning phase. An adaptive trigger generator is
deployed to obtain effective triggers that are difficult to be noticed.
Extensive experiments on real-world datasets against various defense strategies
demonstrate the effectiveness of our proposed method in conducting effective
unnoticeable backdoor attacks.
|
2303.01267 | Lixiang Ru | Lixiang Ru, Heliang Zheng, Yibing Zhan, Bo Du | Token Contrast for Weakly-Supervised Semantic Segmentation | Accepted to CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Weakly-Supervised Semantic Segmentation (WSSS) using image-level labels
typically utilizes Class Activation Map (CAM) to generate the pseudo labels.
Limited by the local structure perception of CNN, CAM usually cannot identify
the integral object regions. Though the recent Vision Transformer (ViT) can
remedy this flaw, we observe it also brings the over-smoothing issue, \ie, the
final patch tokens incline to be uniform. In this work, we propose Token
Contrast (ToCo) to address this issue and further explore the virtue of ViT for
WSSS. Firstly, motivated by the observation that intermediate layers in ViT can
still retain semantic diversity, we designed a Patch Token Contrast module
(PTC). PTC supervises the final patch tokens with the pseudo token relations
derived from intermediate layers, allowing them to align the semantic regions
and thus yield more accurate CAM. Secondly, to further differentiate the
low-confidence regions in CAM, we devised a Class Token Contrast module (CTC)
inspired by the fact that class tokens in ViT can capture high-level semantics.
CTC facilitates the representation consistency between uncertain local regions
and global objects by contrasting their class tokens. Experiments on the PASCAL
VOC and MS COCO datasets show the proposed ToCo can remarkably surpass other
single-stage competitors and achieve comparable performance with
state-of-the-art multi-stage methods. Code is available at
https://github.com/rulixiang/ToCo.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 13:51:58 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Ru",
"Lixiang",
""
],
[
"Zheng",
"Heliang",
""
],
[
"Zhan",
"Yibing",
""
],
[
"Du",
"Bo",
""
]
] | TITLE: Token Contrast for Weakly-Supervised Semantic Segmentation
ABSTRACT: Weakly-Supervised Semantic Segmentation (WSSS) using image-level labels
typically utilizes Class Activation Map (CAM) to generate the pseudo labels.
Limited by the local structure perception of CNN, CAM usually cannot identify
the integral object regions. Though the recent Vision Transformer (ViT) can
remedy this flaw, we observe it also brings the over-smoothing issue, \ie, the
final patch tokens incline to be uniform. In this work, we propose Token
Contrast (ToCo) to address this issue and further explore the virtue of ViT for
WSSS. Firstly, motivated by the observation that intermediate layers in ViT can
still retain semantic diversity, we designed a Patch Token Contrast module
(PTC). PTC supervises the final patch tokens with the pseudo token relations
derived from intermediate layers, allowing them to align the semantic regions
and thus yield more accurate CAM. Secondly, to further differentiate the
low-confidence regions in CAM, we devised a Class Token Contrast module (CTC)
inspired by the fact that class tokens in ViT can capture high-level semantics.
CTC facilitates the representation consistency between uncertain local regions
and global objects by contrasting their class tokens. Experiments on the PASCAL
VOC and MS COCO datasets show the proposed ToCo can remarkably surpass other
single-stage competitors and achieve comparable performance with
state-of-the-art multi-stage methods. Code is available at
https://github.com/rulixiang/ToCo.
|
2303.01268 | Pratinav Seth | Pratinav Seth, Akshat Bhandari and Kumud Lakara | Analyzing Effects of Fake Training Data on the Performance of Deep
Learning Systems | Preprint | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Deep learning models frequently suffer from various problems such as class
imbalance and lack of robustness to distribution shift. It is often difficult
to find data suitable for training beyond the available benchmarks. This is
especially the case for computer vision models. However, with the advent of
Generative Adversarial Networks (GANs), it is now possible to generate
high-quality synthetic data. This synthetic data can be used to alleviate some
of the challenges faced by deep learning models. In this work we present a
detailed analysis of the effect of training computer vision models using
different proportions of synthetic data along with real (organic) data. We
analyze the effect that various quantities of synthetic data, when mixed with
original data, can have on a model's robustness to out-of-distribution data and
the general quality of predictions.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 13:53:22 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Seth",
"Pratinav",
""
],
[
"Bhandari",
"Akshat",
""
],
[
"Lakara",
"Kumud",
""
]
] | TITLE: Analyzing Effects of Fake Training Data on the Performance of Deep
Learning Systems
ABSTRACT: Deep learning models frequently suffer from various problems such as class
imbalance and lack of robustness to distribution shift. It is often difficult
to find data suitable for training beyond the available benchmarks. This is
especially the case for computer vision models. However, with the advent of
Generative Adversarial Networks (GANs), it is now possible to generate
high-quality synthetic data. This synthetic data can be used to alleviate some
of the challenges faced by deep learning models. In this work we present a
detailed analysis of the effect of training computer vision models using
different proportions of synthetic data along with real (organic) data. We
analyze the effect that various quantities of synthetic data, when mixed with
original data, can have on a model's robustness to out-of-distribution data and
the general quality of predictions.
|
2303.01277 | Meng Zhang | Meng Zhang, Qinghao Hu, Peng Sun, Yonggang Wen, Tianwei Zhang | Boosting Distributed Full-graph GNN Training with Asynchronous One-bit
Communication | null | null | null | null | cs.DC cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Training Graph Neural Networks (GNNs) on large graphs is challenging due to
the conflict between the high memory demand and limited GPU memory. Recently,
distributed full-graph GNN training has been widely adopted to tackle this
problem. However, the substantial inter-GPU communication overhead can cause
severe throughput degradation. Existing communication compression techniques
mainly focus on traditional DNN training, whose bottleneck lies in
synchronizing gradients and parameters. We find they do not work well in
distributed GNN training as the barrier is the layer-wise communication of
features during the forward pass & feature gradients during the backward pass.
To this end, we propose an efficient distributed GNN training framework Sylvie,
which employs one-bit quantization technique in GNNs and further pipelines the
curtailed communication with computation to enormously shrink the overhead
while maintaining the model quality. In detail, Sylvie provides a lightweight
Low-bit Module to quantize the sent data and dequantize the received data back
to full precision values in each layer. Additionally, we propose a Bounded
Staleness Adaptor to control the introduced staleness to achieve further
performance enhancement. We conduct theoretical convergence analysis and
extensive experiments on various models & datasets to demonstrate Sylvie can
considerably boost the training throughput by up to 28.1x.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 14:02:39 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Zhang",
"Meng",
""
],
[
"Hu",
"Qinghao",
""
],
[
"Sun",
"Peng",
""
],
[
"Wen",
"Yonggang",
""
],
[
"Zhang",
"Tianwei",
""
]
] | TITLE: Boosting Distributed Full-graph GNN Training with Asynchronous One-bit
Communication
ABSTRACT: Training Graph Neural Networks (GNNs) on large graphs is challenging due to
the conflict between the high memory demand and limited GPU memory. Recently,
distributed full-graph GNN training has been widely adopted to tackle this
problem. However, the substantial inter-GPU communication overhead can cause
severe throughput degradation. Existing communication compression techniques
mainly focus on traditional DNN training, whose bottleneck lies in
synchronizing gradients and parameters. We find they do not work well in
distributed GNN training as the barrier is the layer-wise communication of
features during the forward pass & feature gradients during the backward pass.
To this end, we propose an efficient distributed GNN training framework Sylvie,
which employs one-bit quantization technique in GNNs and further pipelines the
curtailed communication with computation to enormously shrink the overhead
while maintaining the model quality. In detail, Sylvie provides a lightweight
Low-bit Module to quantize the sent data and dequantize the received data back
to full precision values in each layer. Additionally, we propose a Bounded
Staleness Adaptor to control the introduced staleness to achieve further
performance enhancement. We conduct theoretical convergence analysis and
extensive experiments on various models & datasets to demonstrate Sylvie can
considerably boost the training throughput by up to 28.1x.
|
2303.01297 | Raciel Yera | Jes\'us Bobadilla and Abraham Guti\'errez and Raciel Yera and Luis
Mart\'inez | Creating Synthetic Datasets for Collaborative Filtering Recommender
Systems using Generative Adversarial Networks | null | null | null | null | cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Research and education in machine learning needs diverse, representative, and
open datasets that contain sufficient samples to handle the necessary training,
validation, and testing tasks. Currently, the Recommender Systems area includes
a large number of subfields in which accuracy and beyond accuracy quality
measures are continuously improved. To feed this research variety, it is
necessary and convenient to reinforce the existing datasets with synthetic
ones. This paper proposes a Generative Adversarial Network (GAN)-based method
to generate collaborative filtering datasets in a parameterized way, by
selecting their preferred number of users, items, samples, and stochastic
variability. This parameterization cannot be made using regular GANs. Our GAN
model is fed with dense, short, and continuous embedding representations of
items and users, instead of sparse, large, and discrete vectors, to make an
accurate and quick learning, compared to the traditional approach based on
large and sparse input vectors. The proposed architecture includes a DeepMF
model to extract the dense user and item embeddings, as well as a clustering
process to convert from the dense GAN generated samples to the discrete and
sparse ones, necessary to create each required synthetic dataset. The results
of three different source datasets show adequate distributions and expected
quality values and evolutions on the generated datasets compared to the source
ones. Synthetic datasets and source codes are available to researchers.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 14:23:27 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Bobadilla",
"Jesús",
""
],
[
"Gutiérrez",
"Abraham",
""
],
[
"Yera",
"Raciel",
""
],
[
"Martínez",
"Luis",
""
]
] | TITLE: Creating Synthetic Datasets for Collaborative Filtering Recommender
Systems using Generative Adversarial Networks
ABSTRACT: Research and education in machine learning needs diverse, representative, and
open datasets that contain sufficient samples to handle the necessary training,
validation, and testing tasks. Currently, the Recommender Systems area includes
a large number of subfields in which accuracy and beyond accuracy quality
measures are continuously improved. To feed this research variety, it is
necessary and convenient to reinforce the existing datasets with synthetic
ones. This paper proposes a Generative Adversarial Network (GAN)-based method
to generate collaborative filtering datasets in a parameterized way, by
selecting their preferred number of users, items, samples, and stochastic
variability. This parameterization cannot be made using regular GANs. Our GAN
model is fed with dense, short, and continuous embedding representations of
items and users, instead of sparse, large, and discrete vectors, to make an
accurate and quick learning, compared to the traditional approach based on
large and sparse input vectors. The proposed architecture includes a DeepMF
model to extract the dense user and item embeddings, as well as a clustering
process to convert from the dense GAN generated samples to the discrete and
sparse ones, necessary to create each required synthetic dataset. The results
of three different source datasets show adequate distributions and expected
quality values and evolutions on the generated datasets compared to the source
ones. Synthetic datasets and source codes are available to researchers.
|
2303.01341 | Zefa Hu | Zefa Hu, Xiuyi Chen, Haoran Wu, Minglun Han, Ziyi Ni, Jing Shi, Shuang
Xu, Bo Xu | Matching-based Term Semantics Pre-training for Spoken Patient Query
Understanding | ICASSP 2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Medical Slot Filling (MSF) task aims to convert medical queries into
structured information, playing an essential role in diagnosis dialogue
systems. However, the lack of sufficient term semantics learning makes existing
approaches hard to capture semantically identical but colloquial expressions of
terms in medical conversations. In this work, we formalize MSF into a matching
problem and propose a Term Semantics Pre-trained Matching Network (TSPMN) that
takes both terms and queries as input to model their semantic interaction. To
learn term semantics better, we further design two self-supervised objectives,
including Contrastive Term Discrimination (CTD) and Matching-based Mask Term
Modeling (MMTM). CTD determines whether it is the masked term in the dialogue
for each given term, while MMTM directly predicts the masked ones. Experimental
results on two Chinese benchmarks show that TSPMN outperforms strong baselines,
especially in few-shot settings.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 15:17:38 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Hu",
"Zefa",
""
],
[
"Chen",
"Xiuyi",
""
],
[
"Wu",
"Haoran",
""
],
[
"Han",
"Minglun",
""
],
[
"Ni",
"Ziyi",
""
],
[
"Shi",
"Jing",
""
],
[
"Xu",
"Shuang",
""
],
[
"Xu",
"Bo",
""
]
] | TITLE: Matching-based Term Semantics Pre-training for Spoken Patient Query
Understanding
ABSTRACT: Medical Slot Filling (MSF) task aims to convert medical queries into
structured information, playing an essential role in diagnosis dialogue
systems. However, the lack of sufficient term semantics learning makes existing
approaches hard to capture semantically identical but colloquial expressions of
terms in medical conversations. In this work, we formalize MSF into a matching
problem and propose a Term Semantics Pre-trained Matching Network (TSPMN) that
takes both terms and queries as input to model their semantic interaction. To
learn term semantics better, we further design two self-supervised objectives,
including Contrastive Term Discrimination (CTD) and Matching-based Mask Term
Modeling (MMTM). CTD determines whether it is the masked term in the dialogue
for each given term, while MMTM directly predicts the masked ones. Experimental
results on two Chinese benchmarks show that TSPMN outperforms strong baselines,
especially in few-shot settings.
|
2303.01342 | Ario Sadafi | Ario Sadafi, Nassir Navab, Carsten Marr | Active Learning Enhances Classification of Histopathology Whole Slide
Images with Attention-based Multiple Instance Learning | Accepted for publication at the 2023 IEEE International Symposium on
Biomedical Imaging (ISBI 2023) | null | null | null | cs.CV eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In many histopathology tasks, sample classification depends on morphological
details in tissue or single cells that are only visible at the highest
magnification. For a pathologist, this implies tedious zooming in and out,
while for a computational decision support algorithm, it leads to the analysis
of a huge number of small image patches per whole slide image (WSI).
Attention-based multiple instance learning (MIL), where attention estimation is
learned in a weakly supervised manner, has been successfully applied in
computational histopathology, but it is challenged by large numbers of
irrelevant patches, reducing its accuracy. Here, we present an active learning
approach to the problem. Querying the expert to annotate regions of interest in
a WSI guides the formation of high-attention regions for MIL. We train an
attention-based MIL and calculate a confidence metric for every image in the
dataset to select the most uncertain WSIs for expert annotation. We test our
approach on the CAMELYON17 dataset classifying metastatic lymph node sections
in breast cancer. With a novel attention guiding loss, this leads to an
accuracy boost of the trained models with few regions annotated for each class.
Active learning thus improves WSIs classification accuracy, leads to faster and
more robust convergence, and speeds up the annotation process. It may in the
future serve as an important contribution to train MIL models in the clinically
relevant context of cancer classification in histopathology.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 15:18:58 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Sadafi",
"Ario",
""
],
[
"Navab",
"Nassir",
""
],
[
"Marr",
"Carsten",
""
]
] | TITLE: Active Learning Enhances Classification of Histopathology Whole Slide
Images with Attention-based Multiple Instance Learning
ABSTRACT: In many histopathology tasks, sample classification depends on morphological
details in tissue or single cells that are only visible at the highest
magnification. For a pathologist, this implies tedious zooming in and out,
while for a computational decision support algorithm, it leads to the analysis
of a huge number of small image patches per whole slide image (WSI).
Attention-based multiple instance learning (MIL), where attention estimation is
learned in a weakly supervised manner, has been successfully applied in
computational histopathology, but it is challenged by large numbers of
irrelevant patches, reducing its accuracy. Here, we present an active learning
approach to the problem. Querying the expert to annotate regions of interest in
a WSI guides the formation of high-attention regions for MIL. We train an
attention-based MIL and calculate a confidence metric for every image in the
dataset to select the most uncertain WSIs for expert annotation. We test our
approach on the CAMELYON17 dataset classifying metastatic lymph node sections
in breast cancer. With a novel attention guiding loss, this leads to an
accuracy boost of the trained models with few regions annotated for each class.
Active learning thus improves WSIs classification accuracy, leads to faster and
more robust convergence, and speeds up the annotation process. It may in the
future serve as an important contribution to train MIL models in the clinically
relevant context of cancer classification in histopathology.
|
2303.01377 | Daniel Sens | Daniel Sens and Ario Sadafi, Francesco Paolo Casale, Nassir Navab,
Carsten Marr | BEL: A Bag Embedding Loss for Transformer enhances Multiple Instance
Whole Slide Image Classification | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multiple Instance Learning (MIL) has become the predominant approach for
classification tasks on gigapixel histopathology whole slide images (WSIs).
Within the MIL framework, single WSIs (bags) are decomposed into patches
(instances), with only WSI-level annotation available. Recent MIL approaches
produce highly informative bag level representations by utilizing the
transformer architecture's ability to model the dependencies between instances.
However, when applied to high magnification datasets, problems emerge due to
the large number of instances and the weak supervisory learning signal. To
address this problem, we propose to additionally train transformers with a
novel Bag Embedding Loss (BEL). BEL forces the model to learn a discriminative
bag-level representation by minimizing the distance between bag embeddings of
the same class and maximizing the distance between different classes. We
evaluate BEL with the Transformer architecture TransMIL on two publicly
available histopathology datasets, BRACS and CAMELYON17. We show that with BEL,
TransMIL outperforms the baseline models on both datasets, thus contributing to
the clinically highly relevant AI-based tumor classification of histological
patient material.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 16:02:55 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Sens",
"Daniel",
""
],
[
"Sadafi",
"Ario",
""
],
[
"Casale",
"Francesco Paolo",
""
],
[
"Navab",
"Nassir",
""
],
[
"Marr",
"Carsten",
""
]
] | TITLE: BEL: A Bag Embedding Loss for Transformer enhances Multiple Instance
Whole Slide Image Classification
ABSTRACT: Multiple Instance Learning (MIL) has become the predominant approach for
classification tasks on gigapixel histopathology whole slide images (WSIs).
Within the MIL framework, single WSIs (bags) are decomposed into patches
(instances), with only WSI-level annotation available. Recent MIL approaches
produce highly informative bag level representations by utilizing the
transformer architecture's ability to model the dependencies between instances.
However, when applied to high magnification datasets, problems emerge due to
the large number of instances and the weak supervisory learning signal. To
address this problem, we propose to additionally train transformers with a
novel Bag Embedding Loss (BEL). BEL forces the model to learn a discriminative
bag-level representation by minimizing the distance between bag embeddings of
the same class and maximizing the distance between different classes. We
evaluate BEL with the Transformer architecture TransMIL on two publicly
available histopathology datasets, BRACS and CAMELYON17. We show that with BEL,
TransMIL outperforms the baseline models on both datasets, thus contributing to
the clinically highly relevant AI-based tumor classification of histological
patient material.
|
2303.01384 | Benjamin Estermann | Benjamin Estermann and Roger Wattenhofer | DAVA: Disentangling Adversarial Variational Autoencoder | Published as a conference paper at ICLR 2023 | null | null | null | cs.LG cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The use of well-disentangled representations offers many advantages for
downstream tasks, e.g. an increased sample efficiency, or better
interpretability. However, the quality of disentangled interpretations is often
highly dependent on the choice of dataset-specific hyperparameters, in
particular the regularization strength. To address this issue, we introduce
DAVA, a novel training procedure for variational auto-encoders. DAVA completely
alleviates the problem of hyperparameter selection. We compare DAVA to models
with optimal hyperparameters. Without any hyperparameter tuning, DAVA is
competitive on a diverse range of commonly used datasets. Underlying DAVA, we
discover a necessary condition for unsupervised disentanglement, which we call
PIPE. We demonstrate the ability of PIPE to positively predict the performance
of downstream models in abstract reasoning. We also thoroughly investigate
correlations with existing supervised and unsupervised metrics. The code is
available at https://github.com/besterma/dava.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 16:08:23 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Estermann",
"Benjamin",
""
],
[
"Wattenhofer",
"Roger",
""
]
] | TITLE: DAVA: Disentangling Adversarial Variational Autoencoder
ABSTRACT: The use of well-disentangled representations offers many advantages for
downstream tasks, e.g. an increased sample efficiency, or better
interpretability. However, the quality of disentangled interpretations is often
highly dependent on the choice of dataset-specific hyperparameters, in
particular the regularization strength. To address this issue, we introduce
DAVA, a novel training procedure for variational auto-encoders. DAVA completely
alleviates the problem of hyperparameter selection. We compare DAVA to models
with optimal hyperparameters. Without any hyperparameter tuning, DAVA is
competitive on a diverse range of commonly used datasets. Underlying DAVA, we
discover a necessary condition for unsupervised disentanglement, which we call
PIPE. We demonstrate the ability of PIPE to positively predict the performance
of downstream models in abstract reasoning. We also thoroughly investigate
correlations with existing supervised and unsupervised metrics. The code is
available at https://github.com/besterma/dava.
|
2303.01389 | Anna Kurbatskaya | Anna Kurbatskaya, Alberto Jaramillo-Jimenez, John Fredy Ochoa-Gomez,
Kolbj{\o}rn Br{\o}nnick, Alvaro Fernandez-Quilez | Machine Learning-Based Detection of Parkinson's Disease From
Resting-State EEG: A Multi-Center Study | null | null | null | null | eess.SP cs.LG | http://creativecommons.org/licenses/by/4.0/ | Resting-state EEG (rs-EEG) has been demonstrated to aid in Parkinson's
disease (PD) diagnosis. In particular, the power spectral density (PSD) of
low-frequency bands ({\delta} and {\theta}) and high-frequency bands ({\alpha}
and \b{eta}) has been shown to be significantly different in patients with PD
as compared to subjects without PD (non-PD). However, rs-EEG feature extraction
and the interpretation thereof can be time-intensive and prone to examiner
variability. Machine learning (ML) has the potential to automatize the analysis
of rs-EEG recordings and provides a supportive tool for clinicians to ease
their workload. In this work, we use rs-EEG recordings of 84 PD and 85 non-PD
subjects pooled from four datasets obtained at different centers. We propose an
end-to-end pipeline consisting of preprocessing, extraction of PSD features
from clinically validated frequency bands, and feature selection before
evaluating the classification ability of the features via ML algorithms to
stratify between PD and non-PD subjects. Further, we evaluate the effect of
feature harmonization, given the multi-center nature of the datasets. Our
validation results show, on average, an improvement in PD detection ability
(69.6% vs. 75.5% accuracy) by logistic regression when harmonizing the features
and performing univariate feature selection (k = 202 features). Our final
results show an average global accuracy of 72.2% with balanced accuracy results
for all the centers included in the study: 60.6%, 68.7%, 77.7%, and 82.2%,
respectively.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 16:19:24 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Kurbatskaya",
"Anna",
""
],
[
"Jaramillo-Jimenez",
"Alberto",
""
],
[
"Ochoa-Gomez",
"John Fredy",
""
],
[
"Brønnick",
"Kolbjørn",
""
],
[
"Fernandez-Quilez",
"Alvaro",
""
]
] | TITLE: Machine Learning-Based Detection of Parkinson's Disease From
Resting-State EEG: A Multi-Center Study
ABSTRACT: Resting-state EEG (rs-EEG) has been demonstrated to aid in Parkinson's
disease (PD) diagnosis. In particular, the power spectral density (PSD) of
low-frequency bands ({\delta} and {\theta}) and high-frequency bands ({\alpha}
and \b{eta}) has been shown to be significantly different in patients with PD
as compared to subjects without PD (non-PD). However, rs-EEG feature extraction
and the interpretation thereof can be time-intensive and prone to examiner
variability. Machine learning (ML) has the potential to automatize the analysis
of rs-EEG recordings and provides a supportive tool for clinicians to ease
their workload. In this work, we use rs-EEG recordings of 84 PD and 85 non-PD
subjects pooled from four datasets obtained at different centers. We propose an
end-to-end pipeline consisting of preprocessing, extraction of PSD features
from clinically validated frequency bands, and feature selection before
evaluating the classification ability of the features via ML algorithms to
stratify between PD and non-PD subjects. Further, we evaluate the effect of
feature harmonization, given the multi-center nature of the datasets. Our
validation results show, on average, an improvement in PD detection ability
(69.6% vs. 75.5% accuracy) by logistic regression when harmonizing the features
and performing univariate feature selection (k = 202 features). Our final
results show an average global accuracy of 72.2% with balanced accuracy results
for all the centers included in the study: 60.6%, 68.7%, 77.7%, and 82.2%,
respectively.
|
2303.01391 | Hongyao Tang | Hongyao Tang, Min Zhang, Jianye Hao | The Ladder in Chaos: A Simple and Effective Improvement to General DRL
Algorithms by Policy Path Trimming and Boosting | Rudimentary version. Work in progress | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Knowing the learning dynamics of policy is significant to unveiling the
mysteries of Reinforcement Learning (RL). It is especially crucial yet
challenging to Deep RL, from which the remedies to notorious issues like sample
inefficiency and learning instability could be obtained. In this paper, we
study how the policy networks of typical DRL agents evolve during the learning
process by empirically investigating several kinds of temporal change for each
policy parameter. On typical MuJoCo and DeepMind Control Suite (DMC)
benchmarks, we find common phenomena for TD3 and RAD agents: 1) the activity of
policy network parameters is highly asymmetric and policy networks advance
monotonically along very few major parameter directions; 2) severe detours
occur in parameter update and harmonic-like changes are observed for all minor
parameter directions. By performing a novel temporal SVD along policy learning
path, the major and minor parameter directions are identified as the columns of
right unitary matrix associated with dominant and insignificant singular values
respectively. Driven by the discoveries above, we propose a simple and
effective method, called Policy Path Trimming and Boosting (PPTB), as a general
plug-in improvement to DRL algorithms. The key idea of PPTB is to periodically
trim the policy learning path by canceling the policy updates in minor
parameter directions, while boost the learning path by encouraging the advance
in major directions. In experiments, we demonstrate the general and significant
performance improvements brought by PPTB, when combined with TD3 and RAD in
MuJoCo and DMC environments respectively.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 16:20:46 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Tang",
"Hongyao",
""
],
[
"Zhang",
"Min",
""
],
[
"Hao",
"Jianye",
""
]
] | TITLE: The Ladder in Chaos: A Simple and Effective Improvement to General DRL
Algorithms by Policy Path Trimming and Boosting
ABSTRACT: Knowing the learning dynamics of policy is significant to unveiling the
mysteries of Reinforcement Learning (RL). It is especially crucial yet
challenging to Deep RL, from which the remedies to notorious issues like sample
inefficiency and learning instability could be obtained. In this paper, we
study how the policy networks of typical DRL agents evolve during the learning
process by empirically investigating several kinds of temporal change for each
policy parameter. On typical MuJoCo and DeepMind Control Suite (DMC)
benchmarks, we find common phenomena for TD3 and RAD agents: 1) the activity of
policy network parameters is highly asymmetric and policy networks advance
monotonically along very few major parameter directions; 2) severe detours
occur in parameter update and harmonic-like changes are observed for all minor
parameter directions. By performing a novel temporal SVD along policy learning
path, the major and minor parameter directions are identified as the columns of
right unitary matrix associated with dominant and insignificant singular values
respectively. Driven by the discoveries above, we propose a simple and
effective method, called Policy Path Trimming and Boosting (PPTB), as a general
plug-in improvement to DRL algorithms. The key idea of PPTB is to periodically
trim the policy learning path by canceling the policy updates in minor
parameter directions, while boost the learning path by encouraging the advance
in major directions. In experiments, we demonstrate the general and significant
performance improvements brought by PPTB, when combined with TD3 and RAD in
MuJoCo and DMC environments respectively.
|
2303.01412 | Damian Machlanski | Damian Machlanski, Spyridon Samothrakis, Paul Clarke | Hyperparameter Tuning and Model Evaluation in Causal Effect Estimation | 36 pages, 3 figures | null | null | null | cs.LG stat.ME | http://creativecommons.org/licenses/by/4.0/ | The performance of most causal effect estimators relies on accurate
predictions of high-dimensional non-linear functions of the observed data. The
remarkable flexibility of modern Machine Learning (ML) methods is perfectly
suited to this task. However, data-driven hyperparameter tuning of ML methods
requires effective model evaluation to avoid large errors in causal estimates,
a task made more challenging because causal inference involves unavailable
counterfactuals. Multiple performance-validation metrics have recently been
proposed such that practitioners now not only have to make complex decisions
about which causal estimators, ML learners and hyperparameters to choose, but
also about which evaluation metric to use. This paper, motivated by unclear
recommendations, investigates the interplay between the four different aspects
of model evaluation for causal effect estimation. We develop a comprehensive
experimental setup that involves many commonly used causal estimators, ML
methods and evaluation approaches and apply it to four well-known causal
inference benchmark datasets. Our results suggest that optimal hyperparameter
tuning of ML learners is enough to reach state-of-the-art performance in effect
estimation, regardless of estimators and learners. We conclude that most causal
estimators are roughly equivalent in performance if tuned thoroughly enough. We
also find hyperparameter tuning and model evaluation are much more important
than causal estimators and ML methods. Finally, from the significant gap we
find in estimation performance of popular evaluation metrics compared with
optimal model selection choices, we call for more research into causal model
evaluation to unlock the optimum performance not currently being delivered even
by state-of-the-art procedures.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 17:03:02 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Machlanski",
"Damian",
""
],
[
"Samothrakis",
"Spyridon",
""
],
[
"Clarke",
"Paul",
""
]
] | TITLE: Hyperparameter Tuning and Model Evaluation in Causal Effect Estimation
ABSTRACT: The performance of most causal effect estimators relies on accurate
predictions of high-dimensional non-linear functions of the observed data. The
remarkable flexibility of modern Machine Learning (ML) methods is perfectly
suited to this task. However, data-driven hyperparameter tuning of ML methods
requires effective model evaluation to avoid large errors in causal estimates,
a task made more challenging because causal inference involves unavailable
counterfactuals. Multiple performance-validation metrics have recently been
proposed such that practitioners now not only have to make complex decisions
about which causal estimators, ML learners and hyperparameters to choose, but
also about which evaluation metric to use. This paper, motivated by unclear
recommendations, investigates the interplay between the four different aspects
of model evaluation for causal effect estimation. We develop a comprehensive
experimental setup that involves many commonly used causal estimators, ML
methods and evaluation approaches and apply it to four well-known causal
inference benchmark datasets. Our results suggest that optimal hyperparameter
tuning of ML learners is enough to reach state-of-the-art performance in effect
estimation, regardless of estimators and learners. We conclude that most causal
estimators are roughly equivalent in performance if tuned thoroughly enough. We
also find hyperparameter tuning and model evaluation are much more important
than causal estimators and ML methods. Finally, from the significant gap we
find in estimation performance of popular evaluation metrics compared with
optimal model selection choices, we call for more research into causal model
evaluation to unlock the optimum performance not currently being delivered even
by state-of-the-art procedures.
|
2303.01416 | Ivan Skorokhodov | Ivan Skorokhodov, Aliaksandr Siarohin, Yinghao Xu, Jian Ren, Hsin-Ying
Lee, Peter Wonka, Sergey Tulyakov | 3D generation on ImageNet | ICLR 2023 (Oral) | ICLR 2023 | null | null | cs.CV cs.AI cs.GR | http://creativecommons.org/licenses/by/4.0/ | Existing 3D-from-2D generators are typically designed for well-curated
single-category datasets, where all the objects have (approximately) the same
scale, 3D location, and orientation, and the camera always points to the center
of the scene. This makes them inapplicable to diverse, in-the-wild datasets of
non-alignable scenes rendered from arbitrary camera poses. In this work, we
develop a 3D generator with Generic Priors (3DGP): a 3D synthesis framework
with more general assumptions about the training data, and show that it scales
to very challenging datasets, like ImageNet. Our model is based on three new
ideas. First, we incorporate an inaccurate off-the-shelf depth estimator into
3D GAN training via a special depth adaptation module to handle the
imprecision. Then, we create a flexible camera model and a regularization
strategy for it to learn its distribution parameters during training. Finally,
we extend the recent ideas of transferring knowledge from pre-trained
classifiers into GANs for patch-wise trained models by employing a simple
distillation-based technique on top of the discriminator. It achieves more
stable training than the existing methods and speeds up the convergence by at
least 40%. We explore our model on four datasets: SDIP Dogs 256x256, SDIP
Elephants 256x256, LSUN Horses 256x256, and ImageNet 256x256, and demonstrate
that 3DGP outperforms the recent state-of-the-art in terms of both texture and
geometry quality. Code and visualizations:
https://snap-research.github.io/3dgp.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 17:06:57 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Skorokhodov",
"Ivan",
""
],
[
"Siarohin",
"Aliaksandr",
""
],
[
"Xu",
"Yinghao",
""
],
[
"Ren",
"Jian",
""
],
[
"Lee",
"Hsin-Ying",
""
],
[
"Wonka",
"Peter",
""
],
[
"Tulyakov",
"Sergey",
""
]
] | TITLE: 3D generation on ImageNet
ABSTRACT: Existing 3D-from-2D generators are typically designed for well-curated
single-category datasets, where all the objects have (approximately) the same
scale, 3D location, and orientation, and the camera always points to the center
of the scene. This makes them inapplicable to diverse, in-the-wild datasets of
non-alignable scenes rendered from arbitrary camera poses. In this work, we
develop a 3D generator with Generic Priors (3DGP): a 3D synthesis framework
with more general assumptions about the training data, and show that it scales
to very challenging datasets, like ImageNet. Our model is based on three new
ideas. First, we incorporate an inaccurate off-the-shelf depth estimator into
3D GAN training via a special depth adaptation module to handle the
imprecision. Then, we create a flexible camera model and a regularization
strategy for it to learn its distribution parameters during training. Finally,
we extend the recent ideas of transferring knowledge from pre-trained
classifiers into GANs for patch-wise trained models by employing a simple
distillation-based technique on top of the discriminator. It achieves more
stable training than the existing methods and speeds up the convergence by at
least 40%. We explore our model on four datasets: SDIP Dogs 256x256, SDIP
Elephants 256x256, LSUN Horses 256x256, and ImageNet 256x256, and demonstrate
that 3DGP outperforms the recent state-of-the-art in terms of both texture and
geometry quality. Code and visualizations:
https://snap-research.github.io/3dgp.
|
2303.01430 | Mark Roman Miller | Mark Roman Miller (1), Eugy Han (1), Cyan DeVeaux (1), Eliot Jones
(1), Ryan Chen (1), and Jeremy N. Bailenson (1) ((1) Stanford University) | A Large-Scale Study of Personal Identifiability of Virtual Reality
Motion Over Time | 15 pages, 5 figures | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | In recent years, social virtual reality (VR), sometimes described as the
"metaverse," has become widely available. With its potential comes risks,
including risks to privacy. To understand these risks, we study the
identifiability of participants' motion in VR in a dataset of 232 VR users with
eight weekly sessions of about thirty minutes each, totaling 764 hours of
social interaction. The sample is unique as we are able to study the effect of
user, session, and time independently. We find that the number of sessions
recorded greatly increases identifiability, and duration per session increases
identifiability as well, but to a lesser degree. We also find that greater
delay between training and testing sessions reduces identifiability.
Ultimately, understanding the identifiability of VR activities will help
designers, security professionals, and consumer advocates make VR safer.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 17:34:58 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Miller",
"Mark Roman",
"",
"Stanford University"
],
[
"Han",
"Eugy",
"",
"Stanford University"
],
[
"DeVeaux",
"Cyan",
"",
"Stanford University"
],
[
"Jones",
"Eliot",
"",
"Stanford University"
],
[
"Chen",
"Ryan",
"",
"Stanford University"
],
[
"Bailenson",
"Jeremy N.",
"",
"Stanford University"
]
] | TITLE: A Large-Scale Study of Personal Identifiability of Virtual Reality
Motion Over Time
ABSTRACT: In recent years, social virtual reality (VR), sometimes described as the
"metaverse," has become widely available. With its potential comes risks,
including risks to privacy. To understand these risks, we study the
identifiability of participants' motion in VR in a dataset of 232 VR users with
eight weekly sessions of about thirty minutes each, totaling 764 hours of
social interaction. The sample is unique as we are able to study the effect of
user, session, and time independently. We find that the number of sessions
recorded greatly increases identifiability, and duration per session increases
identifiability as well, but to a lesser degree. We also find that greater
delay between training and testing sessions reduces identifiability.
Ultimately, understanding the identifiability of VR activities will help
designers, security professionals, and consumer advocates make VR safer.
|
2303.01435 | Debajyoti Mondal | Subroto Nag Pinku, Debajyoti Mondal and Chanchal K. Roy | Pathways to Leverage Transcompiler based Data Augmentation for
Cross-Language Clone Detection | Accepted at the 31st IEEE/ACM International Conference on Program
Comprehension (ICPC 2023) | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Software clones are often introduced when developers reuse code fragments to
implement similar functionalities in the same or different software systems.
Many high-performing clone detection tools today are based on deep learning
techniques and are mostly used for detecting clones written in the same
programming language, whereas clone detection tools for detecting
cross-language clones are also emerging rapidly. The popularity of deep
learning-based clone detection tools creates an opportunity to investigate how
known strategies that boost the performances of deep learning models could be
further leveraged to improve clone detection tools. In this paper, we
investigate such a strategy, data augmentation, which has not yet been explored
for cross-language clone detection as opposed to single-language clone
detection. We show how the existing knowledge on transcompilers
(source-to-source translators) can be used for data augmentation to boost the
performance of cross-language clone detection models, as well as to adapt
single-language clone detection models to create cross-language clone detection
pipelines. To demonstrate the performance boost for cross-language clone
detection through data augmentation, we exploit Transcoder, which is a
pre-trained source-to-source translator. To show how to extend single-language
models for cross-language clone detection, we extend a popular single-language
model, Graph Matching Network (GMN) in a combination with the transcompilers.
We evaluated our models on popular benchmark datasets. Our experimental results
showed improvements in F1 scores (sometimes up to 3%) for the cutting-edge
cross-language clone detection models. Even when extending GMN for
cross-language clone detection, the models built leveraging data augmentation
outperformed the baseline with scores of 0.90, 0.92, and 0.91 for precision,
recall, and F1 score, respectively.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 17:49:10 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Pinku",
"Subroto Nag",
""
],
[
"Mondal",
"Debajyoti",
""
],
[
"Roy",
"Chanchal K.",
""
]
] | TITLE: Pathways to Leverage Transcompiler based Data Augmentation for
Cross-Language Clone Detection
ABSTRACT: Software clones are often introduced when developers reuse code fragments to
implement similar functionalities in the same or different software systems.
Many high-performing clone detection tools today are based on deep learning
techniques and are mostly used for detecting clones written in the same
programming language, whereas clone detection tools for detecting
cross-language clones are also emerging rapidly. The popularity of deep
learning-based clone detection tools creates an opportunity to investigate how
known strategies that boost the performances of deep learning models could be
further leveraged to improve clone detection tools. In this paper, we
investigate such a strategy, data augmentation, which has not yet been explored
for cross-language clone detection as opposed to single-language clone
detection. We show how the existing knowledge on transcompilers
(source-to-source translators) can be used for data augmentation to boost the
performance of cross-language clone detection models, as well as to adapt
single-language clone detection models to create cross-language clone detection
pipelines. To demonstrate the performance boost for cross-language clone
detection through data augmentation, we exploit Transcoder, which is a
pre-trained source-to-source translator. To show how to extend single-language
models for cross-language clone detection, we extend a popular single-language
model, Graph Matching Network (GMN) in a combination with the transcompilers.
We evaluated our models on popular benchmark datasets. Our experimental results
showed improvements in F1 scores (sometimes up to 3%) for the cutting-edge
cross-language clone detection models. Even when extending GMN for
cross-language clone detection, the models built leveraging data augmentation
outperformed the baseline with scores of 0.90, 0.92, and 0.91 for precision,
recall, and F1 score, respectively.
|
2303.01465 | Anuj Rai | Anuj Rai, Somnath Dey, Pradeep Patidar, Prakhar Rai | MoSFPAD: An end-to-end Ensemble of MobileNet and Support Vector
Classifier for Fingerprint Presentation Attack Detection | 12 pages, 3 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Automatic fingerprint recognition systems are the most extensively used
systems for person authentication although they are vulnerable to Presentation
attacks. Artificial artifacts created with the help of various materials are
used to deceive these systems causing a threat to the security of
fingerprint-based applications. This paper proposes a novel end-to-end model to
detect fingerprint Presentation attacks. The proposed model incorporates
MobileNet as a feature extractor and a Support Vector Classifier as a
classifier to detect presentation attacks in cross-material and cross-sensor
paradigms. The feature extractor's parameters are learned with the loss
generated by the support vector classifier. The proposed model eliminates the
need for intermediary data preparation procedures, unlike other static hybrid
architectures. The performance of the proposed model has been validated on
benchmark LivDet 2011, 2013, 2015, 2017, and 2019 databases, and overall
accuracy of 98.64%, 99.50%, 97.23%, 95.06%, and 95.20% is achieved on these
databases, respectively. The performance of the proposed model is compared with
state-of-the-art methods and the proposed method outperforms in cross-material
and cross-sensor paradigms in terms of average classification error.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:27:48 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Rai",
"Anuj",
""
],
[
"Dey",
"Somnath",
""
],
[
"Patidar",
"Pradeep",
""
],
[
"Rai",
"Prakhar",
""
]
] | TITLE: MoSFPAD: An end-to-end Ensemble of MobileNet and Support Vector
Classifier for Fingerprint Presentation Attack Detection
ABSTRACT: Automatic fingerprint recognition systems are the most extensively used
systems for person authentication although they are vulnerable to Presentation
attacks. Artificial artifacts created with the help of various materials are
used to deceive these systems causing a threat to the security of
fingerprint-based applications. This paper proposes a novel end-to-end model to
detect fingerprint Presentation attacks. The proposed model incorporates
MobileNet as a feature extractor and a Support Vector Classifier as a
classifier to detect presentation attacks in cross-material and cross-sensor
paradigms. The feature extractor's parameters are learned with the loss
generated by the support vector classifier. The proposed model eliminates the
need for intermediary data preparation procedures, unlike other static hybrid
architectures. The performance of the proposed model has been validated on
benchmark LivDet 2011, 2013, 2015, 2017, and 2019 databases, and overall
accuracy of 98.64%, 99.50%, 97.23%, 95.06%, and 95.20% is achieved on these
databases, respectively. The performance of the proposed model is compared with
state-of-the-art methods and the proposed method outperforms in cross-material
and cross-sensor paradigms in terms of average classification error.
|
2303.01475 | Zixuan Liu | Zixuan Liu, Ziqiao Wang, Hongyu Guo, Yongyi Mao | Over-training with Mixup May Hurt Generalization | Accepted to ICLR 2023 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mixup, which creates synthetic training instances by linearly interpolating
random sample pairs, is a simple and yet effective regularization technique to
boost the performance of deep models trained with SGD. In this work, we report
a previously unobserved phenomenon in Mixup training: on a number of standard
datasets, the performance of Mixup-trained models starts to decay after
training for a large number of epochs, giving rise to a U-shaped generalization
curve. This behavior is further aggravated when the size of original dataset is
reduced. To help understand such a behavior of Mixup, we show theoretically
that Mixup training may introduce undesired data-dependent label noises to the
synthesized data. Via analyzing a least-square regression problem with a random
feature model, we explain why noisy labels may cause the U-shaped curve to
occur: Mixup improves generalization through fitting the clean patterns at the
early training stage, but as training progresses, Mixup becomes over-fitting to
the noise in the synthetic data. Extensive experiments are performed on a
variety of benchmark datasets, validating this explanation.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:37:34 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Liu",
"Zixuan",
""
],
[
"Wang",
"Ziqiao",
""
],
[
"Guo",
"Hongyu",
""
],
[
"Mao",
"Yongyi",
""
]
] | TITLE: Over-training with Mixup May Hurt Generalization
ABSTRACT: Mixup, which creates synthetic training instances by linearly interpolating
random sample pairs, is a simple and yet effective regularization technique to
boost the performance of deep models trained with SGD. In this work, we report
a previously unobserved phenomenon in Mixup training: on a number of standard
datasets, the performance of Mixup-trained models starts to decay after
training for a large number of epochs, giving rise to a U-shaped generalization
curve. This behavior is further aggravated when the size of original dataset is
reduced. To help understand such a behavior of Mixup, we show theoretically
that Mixup training may introduce undesired data-dependent label noises to the
synthesized data. Via analyzing a least-square regression problem with a random
feature model, we explain why noisy labels may cause the U-shaped curve to
occur: Mixup improves generalization through fitting the clean patterns at the
early training stage, but as training progresses, Mixup becomes over-fitting to
the noise in the synthetic data. Extensive experiments are performed on a
variety of benchmark datasets, validating this explanation.
|
2303.01480 | Jiaming Zhang | Jiaming Zhang, Ruiping Liu, Hao Shi, Kailun Yang, Simon Rei{\ss},
Kunyu Peng, Haodong Fu, Kaiwei Wang, Rainer Stiefelhagen | Delivering Arbitrary-Modal Semantic Segmentation | Accepted by CVPR 2023. Dataset and our code are at:
https://jamycheung.github.io/DELIVER.html | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal fusion can make semantic segmentation more robust. However, fusing
an arbitrary number of modalities remains underexplored. To delve into this
problem, we create the DeLiVER arbitrary-modal segmentation benchmark, covering
Depth, LiDAR, multiple Views, Events, and RGB. Aside from this, we provide this
dataset in four severe weather conditions as well as five sensor failure cases
to exploit modal complementarity and resolve partial outages. To make this
possible, we present the arbitrary cross-modal segmentation model CMNeXt. It
encompasses a Self-Query Hub (SQ-Hub) designed to extract effective information
from any modality for subsequent fusion with the RGB representation and adds
only negligible amounts of parameters (~0.01M) per additional modality. On top,
to efficiently and flexibly harvest discriminative cues from the auxiliary
modalities, we introduce the simple Parallel Pooling Mixer (PPX). With
extensive experiments on a total of six benchmarks, our CMNeXt achieves
state-of-the-art performance on the DeLiVER, KITTI-360, MFNet, NYU Depth V2,
UrbanLF, and MCubeS datasets, allowing to scale from 1 to 81 modalities. On the
freshly collected DeLiVER, the quad-modal CMNeXt reaches up to 66.30% in mIoU
with a +9.10% gain as compared to the mono-modal baseline. The DeLiVER dataset
and our code are at: https://jamycheung.github.io/DELIVER.html.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:41:41 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Zhang",
"Jiaming",
""
],
[
"Liu",
"Ruiping",
""
],
[
"Shi",
"Hao",
""
],
[
"Yang",
"Kailun",
""
],
[
"Reiß",
"Simon",
""
],
[
"Peng",
"Kunyu",
""
],
[
"Fu",
"Haodong",
""
],
[
"Wang",
"Kaiwei",
""
],
[
"Stiefelhagen",
"Rainer",
""
]
] | TITLE: Delivering Arbitrary-Modal Semantic Segmentation
ABSTRACT: Multimodal fusion can make semantic segmentation more robust. However, fusing
an arbitrary number of modalities remains underexplored. To delve into this
problem, we create the DeLiVER arbitrary-modal segmentation benchmark, covering
Depth, LiDAR, multiple Views, Events, and RGB. Aside from this, we provide this
dataset in four severe weather conditions as well as five sensor failure cases
to exploit modal complementarity and resolve partial outages. To make this
possible, we present the arbitrary cross-modal segmentation model CMNeXt. It
encompasses a Self-Query Hub (SQ-Hub) designed to extract effective information
from any modality for subsequent fusion with the RGB representation and adds
only negligible amounts of parameters (~0.01M) per additional modality. On top,
to efficiently and flexibly harvest discriminative cues from the auxiliary
modalities, we introduce the simple Parallel Pooling Mixer (PPX). With
extensive experiments on a total of six benchmarks, our CMNeXt achieves
state-of-the-art performance on the DeLiVER, KITTI-360, MFNet, NYU Depth V2,
UrbanLF, and MCubeS datasets, allowing to scale from 1 to 81 modalities. On the
freshly collected DeLiVER, the quad-modal CMNeXt reaches up to 66.30% in mIoU
with a +9.10% gain as compared to the mono-modal baseline. The DeLiVER dataset
and our code are at: https://jamycheung.github.io/DELIVER.html.
|
2303.01487 | Hasini Witharana | Hasini Witharana, Daniel Volya and Prabhat Mishra | quAssert: Automatic Generation of Quantum Assertions | null | null | null | null | quant-ph cs.SE | http://creativecommons.org/licenses/by/4.0/ | Functional validation is necessary to detect any errors during quantum
computation. There are promising avenues to debug quantum circuits using
runtime assertions. However, the existing approaches rely on the expertise of
the verification engineers to manually design and insert the assertions in
suitable locations. In this paper, we propose automated generation and
placement of quantum assertions based on static analysis and random sampling of
quantum circuits. Specifically, this paper makes two important contributions.
We automatically uncover special properties of a quantum circuit, such as
purely classical states, superposition states, and entangled states using
statistical methods. We also perform automated placement of quantum assertions
to maximize the functional coverage as well as minimize the hardware overhead.
We demonstrate the effectiveness of the generated assertions in error detection
using a suite of quantum benchmarks, including Shor's factoring algorithm and
Grover's search algorithm.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:49:14 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Witharana",
"Hasini",
""
],
[
"Volya",
"Daniel",
""
],
[
"Mishra",
"Prabhat",
""
]
] | TITLE: quAssert: Automatic Generation of Quantum Assertions
ABSTRACT: Functional validation is necessary to detect any errors during quantum
computation. There are promising avenues to debug quantum circuits using
runtime assertions. However, the existing approaches rely on the expertise of
the verification engineers to manually design and insert the assertions in
suitable locations. In this paper, we propose automated generation and
placement of quantum assertions based on static analysis and random sampling of
quantum circuits. Specifically, this paper makes two important contributions.
We automatically uncover special properties of a quantum circuit, such as
purely classical states, superposition states, and entangled states using
statistical methods. We also perform automated placement of quantum assertions
to maximize the functional coverage as well as minimize the hardware overhead.
We demonstrate the effectiveness of the generated assertions in error detection
using a suite of quantum benchmarks, including Shor's factoring algorithm and
Grover's search algorithm.
|
2303.01488 | Archit Sharma | Archit Sharma, Ahmed M. Ahmed, Rehaan Ahmad, Chelsea Finn | Self-Improving Robots: End-to-End Autonomous Visuomotor Reinforcement
Learning | null | null | null | null | cs.RO cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | In imitation and reinforcement learning, the cost of human supervision limits
the amount of data that robots can be trained on. An aspirational goal is to
construct self-improving robots: robots that can learn and improve on their
own, from autonomous interaction with minimal human supervision or oversight.
Such robots could collect and train on much larger datasets, and thus learn
more robust and performant policies. While reinforcement learning offers a
framework for such autonomous learning via trial-and-error, practical
realizations end up requiring extensive human supervision for reward function
design and repeated resetting of the environment between episodes of
interactions. In this work, we propose MEDAL++, a novel design for
self-improving robotic systems: given a small set of expert demonstrations at
the start, the robot autonomously practices the task by learning to both do and
undo the task, simultaneously inferring the reward function from the
demonstrations. The policy and reward function are learned end-to-end from
high-dimensional visual inputs, bypassing the need for explicit state
estimation or task-specific pre-training for visual encoders used in prior
work. We first evaluate our proposed algorithm on a simulated non-episodic
benchmark EARL, finding that MEDAL++ is both more data efficient and gets up to
30% better final performance compared to state-of-the-art vision-based methods.
Our real-robot experiments show that MEDAL++ can be applied to manipulation
problems in larger environments than those considered in prior work, and
autonomous self-improvement can improve the success rate by 30-70% over
behavior cloning on just the expert data. Code, training and evaluation videos
along with a brief overview is available at:
https://architsharma97.github.io/self-improving-robots/
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:51:38 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Sharma",
"Archit",
""
],
[
"Ahmed",
"Ahmed M.",
""
],
[
"Ahmad",
"Rehaan",
""
],
[
"Finn",
"Chelsea",
""
]
] | TITLE: Self-Improving Robots: End-to-End Autonomous Visuomotor Reinforcement
Learning
ABSTRACT: In imitation and reinforcement learning, the cost of human supervision limits
the amount of data that robots can be trained on. An aspirational goal is to
construct self-improving robots: robots that can learn and improve on their
own, from autonomous interaction with minimal human supervision or oversight.
Such robots could collect and train on much larger datasets, and thus learn
more robust and performant policies. While reinforcement learning offers a
framework for such autonomous learning via trial-and-error, practical
realizations end up requiring extensive human supervision for reward function
design and repeated resetting of the environment between episodes of
interactions. In this work, we propose MEDAL++, a novel design for
self-improving robotic systems: given a small set of expert demonstrations at
the start, the robot autonomously practices the task by learning to both do and
undo the task, simultaneously inferring the reward function from the
demonstrations. The policy and reward function are learned end-to-end from
high-dimensional visual inputs, bypassing the need for explicit state
estimation or task-specific pre-training for visual encoders used in prior
work. We first evaluate our proposed algorithm on a simulated non-episodic
benchmark EARL, finding that MEDAL++ is both more data efficient and gets up to
30% better final performance compared to state-of-the-art vision-based methods.
Our real-robot experiments show that MEDAL++ can be applied to manipulation
problems in larger environments than those considered in prior work, and
autonomous self-improvement can improve the success rate by 30-70% over
behavior cloning on just the expert data. Code, training and evaluation videos
along with a brief overview is available at:
https://architsharma97.github.io/self-improving-robots/
|
2303.01490 | Kai North | Marcos Zampieri, Kai North, Tommi Jauhiainen, Mariano Felice, Neha
Kumari, Nishant Nair, Yash Bangera | Language Variety Identification with True Labels | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Language identification is an important first step in many IR and NLP
applications. Most publicly available language identification datasets,
however, are compiled under the assumption that the gold label of each instance
is determined by where texts are retrieved from. Research has shown that this
is a problematic assumption, particularly in the case of very similar languages
(e.g., Croatian and Serbian) and national language varieties (e.g., Brazilian
and European Portuguese), where texts may contain no distinctive marker of the
particular language or variety. To overcome this important limitation, this
paper presents DSL True Labels (DSL-TL), the first human-annotated multilingual
dataset for language variety identification. DSL-TL contains a total of 12,900
instances in Portuguese, split between European Portuguese and Brazilian
Portuguese; Spanish, split between Argentine Spanish and Castilian Spanish; and
English, split between American English and British English. We trained
multiple models to discriminate between these language varieties, and we
present the results in detail. The data and models presented in this paper
provide a reliable benchmark toward the development of robust and fairer
language variety identification systems. We make DSL-TL freely available to the
research community.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:51:58 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Zampieri",
"Marcos",
""
],
[
"North",
"Kai",
""
],
[
"Jauhiainen",
"Tommi",
""
],
[
"Felice",
"Mariano",
""
],
[
"Kumari",
"Neha",
""
],
[
"Nair",
"Nishant",
""
],
[
"Bangera",
"Yash",
""
]
] | TITLE: Language Variety Identification with True Labels
ABSTRACT: Language identification is an important first step in many IR and NLP
applications. Most publicly available language identification datasets,
however, are compiled under the assumption that the gold label of each instance
is determined by where texts are retrieved from. Research has shown that this
is a problematic assumption, particularly in the case of very similar languages
(e.g., Croatian and Serbian) and national language varieties (e.g., Brazilian
and European Portuguese), where texts may contain no distinctive marker of the
particular language or variety. To overcome this important limitation, this
paper presents DSL True Labels (DSL-TL), the first human-annotated multilingual
dataset for language variety identification. DSL-TL contains a total of 12,900
instances in Portuguese, split between European Portuguese and Brazilian
Portuguese; Spanish, split between Argentine Spanish and Castilian Spanish; and
English, split between American English and British English. We trained
multiple models to discriminate between these language varieties, and we
present the results in detail. The data and models presented in this paper
provide a reliable benchmark toward the development of robust and fairer
language variety identification systems. We make DSL-TL freely available to the
research community.
|
2303.01491 | Umang Gupta | Umang Gupta, Tamoghna Chattopadhyay, Nikhil Dhinagar, Paul M.
Thompson, Greg Ver Steeg, The Alzheimer's Disease Neuroimaging Initiative
(ADNI) | Transferring Models Trained on Natural Images to 3D MRI via Position
Encoded Slice Models | To appear at IEEE International Symposium on Biomedical Imaging 2023
(ISBI 2023). Code is available at
https://github.com/umgupta/2d-slice-set-networks | null | null | null | eess.IV cs.LG q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Transfer learning has remarkably improved computer vision. These advances
also promise improvements in neuroimaging, where training set sizes are often
small. However, various difficulties arise in directly applying models
pretrained on natural images to radiologic images, such as MRIs. In particular,
a mismatch in the input space (2D images vs. 3D MRIs) restricts the direct
transfer of models, often forcing us to consider only a few MRI slices as
input. To this end, we leverage the 2D-Slice-CNN architecture of Gupta et al.
(2021), which embeds all the MRI slices with 2D encoders (neural networks that
take 2D image input) and combines them via permutation-invariant layers. With
the insight that the pretrained model can serve as the 2D encoder, we
initialize the 2D encoder with ImageNet pretrained weights that outperform
those initialized and trained from scratch on two neuroimaging tasks -- brain
age prediction on the UK Biobank dataset and Alzheimer's disease detection on
the ADNI dataset. Further, we improve the modeling capabilities of 2D-Slice
models by incorporating spatial information through position embeddings, which
can improve the performance in some cases.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:52:31 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Gupta",
"Umang",
"",
"ADNI"
],
[
"Chattopadhyay",
"Tamoghna",
"",
"ADNI"
],
[
"Dhinagar",
"Nikhil",
"",
"ADNI"
],
[
"Thompson",
"Paul M.",
"",
"ADNI"
],
[
"Steeg",
"Greg Ver",
"",
"ADNI"
],
[
"Initiative",
"The Alzheimer's Disease Neuroimaging",
"",
"ADNI"
]
] | TITLE: Transferring Models Trained on Natural Images to 3D MRI via Position
Encoded Slice Models
ABSTRACT: Transfer learning has remarkably improved computer vision. These advances
also promise improvements in neuroimaging, where training set sizes are often
small. However, various difficulties arise in directly applying models
pretrained on natural images to radiologic images, such as MRIs. In particular,
a mismatch in the input space (2D images vs. 3D MRIs) restricts the direct
transfer of models, often forcing us to consider only a few MRI slices as
input. To this end, we leverage the 2D-Slice-CNN architecture of Gupta et al.
(2021), which embeds all the MRI slices with 2D encoders (neural networks that
take 2D image input) and combines them via permutation-invariant layers. With
the insight that the pretrained model can serve as the 2D encoder, we
initialize the 2D encoder with ImageNet pretrained weights that outperform
those initialized and trained from scratch on two neuroimaging tasks -- brain
age prediction on the UK Biobank dataset and Alzheimer's disease detection on
the ADNI dataset. Further, we improve the modeling capabilities of 2D-Slice
models by incorporating spatial information through position embeddings, which
can improve the performance in some cases.
|
2303.01494 | Xu Ma | Xu Ma, Yuqian Zhou, Huan Wang, Can Qin, Bin Sun, Chang Liu, Yun Fu | Image as Set of Points | ICLR'23 Oral (top 5%); Codes:
https://github.com/ma-xu/Context-Cluster | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | What is an image and how to extract latent features? Convolutional Networks
(ConvNets) consider an image as organized pixels in a rectangular shape and
extract features via convolutional operation in local region; Vision
Transformers (ViTs) treat an image as a sequence of patches and extract
features via attention mechanism in a global range. In this work, we introduce
a straightforward and promising paradigm for visual representation, which is
called Context Clusters. Context clusters (CoCs) view an image as a set of
unorganized points and extract features via simplified clustering algorithm. In
detail, each point includes the raw feature (e.g., color) and positional
information (e.g., coordinates), and a simplified clustering algorithm is
employed to group and extract deep features hierarchically. Our CoCs are
convolution- and attention-free, and only rely on clustering algorithm for
spatial interaction. Owing to the simple design, we show CoCs endow gratifying
interpretability via the visualization of clustering process. Our CoCs aim at
providing a new perspective on image and visual representation, which may enjoy
broad applications in different domains and exhibit profound insights. Even
though we are not targeting SOTA performance, COCs still achieve comparable or
even better results than ConvNets or ViTs on several benchmarks. Codes are
available at: https://github.com/ma-xu/Context-Cluster.
| [
{
"version": "v1",
"created": "Thu, 2 Mar 2023 18:56:39 GMT"
}
] | 2023-03-03T00:00:00 | [
[
"Ma",
"Xu",
""
],
[
"Zhou",
"Yuqian",
""
],
[
"Wang",
"Huan",
""
],
[
"Qin",
"Can",
""
],
[
"Sun",
"Bin",
""
],
[
"Liu",
"Chang",
""
],
[
"Fu",
"Yun",
""
]
] | TITLE: Image as Set of Points
ABSTRACT: What is an image and how to extract latent features? Convolutional Networks
(ConvNets) consider an image as organized pixels in a rectangular shape and
extract features via convolutional operation in local region; Vision
Transformers (ViTs) treat an image as a sequence of patches and extract
features via attention mechanism in a global range. In this work, we introduce
a straightforward and promising paradigm for visual representation, which is
called Context Clusters. Context clusters (CoCs) view an image as a set of
unorganized points and extract features via simplified clustering algorithm. In
detail, each point includes the raw feature (e.g., color) and positional
information (e.g., coordinates), and a simplified clustering algorithm is
employed to group and extract deep features hierarchically. Our CoCs are
convolution- and attention-free, and only rely on clustering algorithm for
spatial interaction. Owing to the simple design, we show CoCs endow gratifying
interpretability via the visualization of clustering process. Our CoCs aim at
providing a new perspective on image and visual representation, which may enjoy
broad applications in different domains and exhibit profound insights. Even
though we are not targeting SOTA performance, COCs still achieve comparable or
even better results than ConvNets or ViTs on several benchmarks. Codes are
available at: https://github.com/ma-xu/Context-Cluster.
|
1910.02684 | Zhou Ziang | Ziang Zhou, Jieming Shi, Shengzhong Zhang, Zengfeng Huang, Qing Li | Effective Stabilized Self-Training on Few-Labeled Graph Data | 34 pages | Zhou, Z., Shi, J., Zhang, S., Huang, Z., & Li, Q. (2023).
Effective Stabilized Self-Training on Few-Labeled Graph Data. Information
Sciences | 10.1016/j.ins.2023.02.032 | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph neural networks (GNNs) are designed for semi-supervised node
classification on graphs where only a subset of nodes have class labels.
However, under extreme cases when very few labels are available (e.g., 1
labeled node per class), GNNs suffer from severe performance degradation.
Specifically, we observe that existing GNNs suffer from unstable training
process on few-labeled graphs, resulting to inferior performance on node
classification. Therefore, we propose an effective framework, Stabilized
Self-Training (SST), which is applicable to existing GNNs to handle the
scarcity of labeled data, and consequently, boost classification accuracy. We
conduct thorough empirical and theoretical analysis to support our findings and
motivate the algorithmic designs in SST. We apply SST to two popular GNN models
GCN and DAGNN, to get SSTGCN and SSTDA methods respectively, and evaluate the
two methods against 10 competitors over 5 benchmarking datasets. Extensive
experiments show that the proposed SST framework is highly effective,
especially when few labeled data are available. Our methods achieve superior
performance under almost all settings over all datasets. For instance, on a
Cora dataset with only 1 labeled node per class, the accuracy of SSTGCN is
62.5%, 17.9% higher than GCN, and the accuracy of SSTDA is 66.4%, which
outperforms DAGNN by 6.6%.
| [
{
"version": "v1",
"created": "Mon, 7 Oct 2019 09:21:49 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Jun 2021 06:40:08 GMT"
},
{
"version": "v3",
"created": "Sat, 9 Apr 2022 15:28:03 GMT"
},
{
"version": "v4",
"created": "Wed, 1 Mar 2023 13:12:23 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Zhou",
"Ziang",
""
],
[
"Shi",
"Jieming",
""
],
[
"Zhang",
"Shengzhong",
""
],
[
"Huang",
"Zengfeng",
""
],
[
"Li",
"Qing",
""
]
] | TITLE: Effective Stabilized Self-Training on Few-Labeled Graph Data
ABSTRACT: Graph neural networks (GNNs) are designed for semi-supervised node
classification on graphs where only a subset of nodes have class labels.
However, under extreme cases when very few labels are available (e.g., 1
labeled node per class), GNNs suffer from severe performance degradation.
Specifically, we observe that existing GNNs suffer from unstable training
process on few-labeled graphs, resulting to inferior performance on node
classification. Therefore, we propose an effective framework, Stabilized
Self-Training (SST), which is applicable to existing GNNs to handle the
scarcity of labeled data, and consequently, boost classification accuracy. We
conduct thorough empirical and theoretical analysis to support our findings and
motivate the algorithmic designs in SST. We apply SST to two popular GNN models
GCN and DAGNN, to get SSTGCN and SSTDA methods respectively, and evaluate the
two methods against 10 competitors over 5 benchmarking datasets. Extensive
experiments show that the proposed SST framework is highly effective,
especially when few labeled data are available. Our methods achieve superior
performance under almost all settings over all datasets. For instance, on a
Cora dataset with only 1 labeled node per class, the accuracy of SSTGCN is
62.5%, 17.9% higher than GCN, and the accuracy of SSTDA is 66.4%, which
outperforms DAGNN by 6.6%.
|
2006.05534 | Chieh-Hsin Lai | Chieh-Hsin Lai, Dongmian Zou and Gilad Lerman | Novelty Detection via Robust Variational Autoencoding | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a new method for novelty detection that can tolerate high
corruption of the training points, whereas previous works assumed either no or
very low corruption. Our method trains a robust variational autoencoder (VAE),
which aims to generate a model for the uncorrupted training points. To gain
robustness to high corruption, we incorporate the following four changes to the
common VAE: 1. Extracting crucial features of the latent code by a carefully
designed dimension reduction component for distributions; 2. Modeling the
latent distribution as a mixture of Gaussian low-rank inliers and full-rank
outliers, where the testing only uses the inlier model; 3. Applying the
Wasserstein-1 metric for regularization, instead of the Kullback-Leibler (KL)
divergence; and 4. Using a robust error for reconstruction. We establish both
robustness to outliers and suitability to low-rank modeling of the Wasserstein
metric as opposed to the KL divergence. We illustrate state-of-the-art results
on standard benchmarks.
| [
{
"version": "v1",
"created": "Tue, 9 Jun 2020 22:34:27 GMT"
},
{
"version": "v2",
"created": "Tue, 6 Oct 2020 14:03:06 GMT"
},
{
"version": "v3",
"created": "Wed, 7 Oct 2020 00:56:34 GMT"
},
{
"version": "v4",
"created": "Wed, 1 Mar 2023 09:03:21 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Lai",
"Chieh-Hsin",
""
],
[
"Zou",
"Dongmian",
""
],
[
"Lerman",
"Gilad",
""
]
] | TITLE: Novelty Detection via Robust Variational Autoencoding
ABSTRACT: We propose a new method for novelty detection that can tolerate high
corruption of the training points, whereas previous works assumed either no or
very low corruption. Our method trains a robust variational autoencoder (VAE),
which aims to generate a model for the uncorrupted training points. To gain
robustness to high corruption, we incorporate the following four changes to the
common VAE: 1. Extracting crucial features of the latent code by a carefully
designed dimension reduction component for distributions; 2. Modeling the
latent distribution as a mixture of Gaussian low-rank inliers and full-rank
outliers, where the testing only uses the inlier model; 3. Applying the
Wasserstein-1 metric for regularization, instead of the Kullback-Leibler (KL)
divergence; and 4. Using a robust error for reconstruction. We establish both
robustness to outliers and suitability to low-rank modeling of the Wasserstein
metric as opposed to the KL divergence. We illustrate state-of-the-art results
on standard benchmarks.
|
2012.08950 | Chang Liu | Chang Liu, Zetian Jiang, Runzhong Wang, Junchi Yan, Lingxiao Huang,
Pinyan Lu | Revocable Deep Reinforcement Learning with Affinity Regularization for
Outlier-Robust Graph Matching | Proceedings of The Eleventh International Conference on Learning
Representations (ICLR 2023) | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph matching (GM) has been a building block in various areas including
computer vision and pattern recognition. Despite recent impressive progress,
existing deep GM methods often have obvious difficulty in handling outliers,
which are ubiquitous in practice. We propose a deep reinforcement learning
based approach RGM, whose sequential node matching scheme naturally fits the
strategy for selective inlier matching against outliers. A revocable action
framework is devised to improve the agent's flexibility against the complex
constrained GM. Moreover, we propose a quadratic approximation technique to
regularize the affinity score, in the presence of outliers. As such, the agent
can finish inlier matching timely when the affinity score stops growing, for
which otherwise an additional parameter i.e. the number of inliers is needed to
avoid matching outliers. In this paper, we focus on learning the back-end
solver under the most general form of GM: the Lawler's QAP, whose input is the
affinity matrix. Especially, our approach can also boost existing GM methods
that use such input. Experiments on multiple real-world datasets demonstrate
its performance regarding both accuracy and robustness.
| [
{
"version": "v1",
"created": "Wed, 16 Dec 2020 13:48:48 GMT"
},
{
"version": "v2",
"created": "Thu, 18 Mar 2021 07:26:13 GMT"
},
{
"version": "v3",
"created": "Tue, 17 Aug 2021 08:58:44 GMT"
},
{
"version": "v4",
"created": "Sun, 14 Aug 2022 02:43:57 GMT"
},
{
"version": "v5",
"created": "Wed, 1 Mar 2023 07:35:43 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Liu",
"Chang",
""
],
[
"Jiang",
"Zetian",
""
],
[
"Wang",
"Runzhong",
""
],
[
"Yan",
"Junchi",
""
],
[
"Huang",
"Lingxiao",
""
],
[
"Lu",
"Pinyan",
""
]
] | TITLE: Revocable Deep Reinforcement Learning with Affinity Regularization for
Outlier-Robust Graph Matching
ABSTRACT: Graph matching (GM) has been a building block in various areas including
computer vision and pattern recognition. Despite recent impressive progress,
existing deep GM methods often have obvious difficulty in handling outliers,
which are ubiquitous in practice. We propose a deep reinforcement learning
based approach RGM, whose sequential node matching scheme naturally fits the
strategy for selective inlier matching against outliers. A revocable action
framework is devised to improve the agent's flexibility against the complex
constrained GM. Moreover, we propose a quadratic approximation technique to
regularize the affinity score, in the presence of outliers. As such, the agent
can finish inlier matching timely when the affinity score stops growing, for
which otherwise an additional parameter i.e. the number of inliers is needed to
avoid matching outliers. In this paper, we focus on learning the back-end
solver under the most general form of GM: the Lawler's QAP, whose input is the
affinity matrix. Especially, our approach can also boost existing GM methods
that use such input. Experiments on multiple real-world datasets demonstrate
its performance regarding both accuracy and robustness.
|
2102.10343 | Sizhe Chen | Sizhe Chen, Qinghua Tao, Zhixing Ye, Xiaolin Huang | Measuring the Transferability of $\ell_\infty$ Attacks by the $\ell_2$
Norm | ICASSP 2023 | null | null | null | cs.LG cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks could be fooled by adversarial examples with trivial
differences to original samples. To keep the difference imperceptible in human
eyes, researchers bound the adversarial perturbations by the $\ell_\infty$
norm, which is now commonly served as the standard to align the strength of
different attacks for a fair comparison. However, we propose that using the
$\ell_\infty$ norm alone is not sufficient in measuring the attack strength,
because even with a fixed $\ell_\infty$ distance, the $\ell_2$ distance also
greatly affects the attack transferability between models. Through the
discovery, we reach more in-depth understandings towards the attack mechanism,
i.e., several existing methods attack black-box models better partly because
they craft perturbations with 70% to 130% larger $\ell_2$ distances. Since
larger perturbations naturally lead to better transferability, we thereby
advocate that the strength of attacks should be simultaneously measured by both
the $\ell_\infty$ and $\ell_2$ norm. Our proposal is firmly supported by
extensive experiments on ImageNet dataset from 7 attacks, 4 white-box models,
and 9 black-box models.
| [
{
"version": "v1",
"created": "Sat, 20 Feb 2021 13:19:31 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Sep 2021 01:40:06 GMT"
},
{
"version": "v3",
"created": "Thu, 17 Feb 2022 04:37:59 GMT"
},
{
"version": "v4",
"created": "Wed, 1 Mar 2023 13:05:43 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Chen",
"Sizhe",
""
],
[
"Tao",
"Qinghua",
""
],
[
"Ye",
"Zhixing",
""
],
[
"Huang",
"Xiaolin",
""
]
] | TITLE: Measuring the Transferability of $\ell_\infty$ Attacks by the $\ell_2$
Norm
ABSTRACT: Deep neural networks could be fooled by adversarial examples with trivial
differences to original samples. To keep the difference imperceptible in human
eyes, researchers bound the adversarial perturbations by the $\ell_\infty$
norm, which is now commonly served as the standard to align the strength of
different attacks for a fair comparison. However, we propose that using the
$\ell_\infty$ norm alone is not sufficient in measuring the attack strength,
because even with a fixed $\ell_\infty$ distance, the $\ell_2$ distance also
greatly affects the attack transferability between models. Through the
discovery, we reach more in-depth understandings towards the attack mechanism,
i.e., several existing methods attack black-box models better partly because
they craft perturbations with 70% to 130% larger $\ell_2$ distances. Since
larger perturbations naturally lead to better transferability, we thereby
advocate that the strength of attacks should be simultaneously measured by both
the $\ell_\infty$ and $\ell_2$ norm. Our proposal is firmly supported by
extensive experiments on ImageNet dataset from 7 attacks, 4 white-box models,
and 9 black-box models.
|
2107.00372 | Jianing Qiu | Jianing Qiu, Frank P.-W. Lo, Xiao Gu, Modou L. Jobarteh, Wenyan Jia,
Tom Baranowski, Matilda Steiner-Asiedu, Alex K. Anderson, Megan A McCrory,
Edward Sazonov, Mingui Sun, Gary Frost, Benny Lo | Egocentric Image Captioning for Privacy-Preserved Passive Dietary Intake
Monitoring | null | IEEE Transactions on Cybernetics, 2023 | 10.1109/TCYB.2023.3243999 | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Camera-based passive dietary intake monitoring is able to continuously
capture the eating episodes of a subject, recording rich visual information,
such as the type and volume of food being consumed, as well as the eating
behaviours of the subject. However, there currently is no method that is able
to incorporate these visual clues and provide a comprehensive context of
dietary intake from passive recording (e.g., is the subject sharing food with
others, what food the subject is eating, and how much food is left in the
bowl). On the other hand, privacy is a major concern while egocentric wearable
cameras are used for capturing. In this paper, we propose a privacy-preserved
secure solution (i.e., egocentric image captioning) for dietary assessment with
passive monitoring, which unifies food recognition, volume estimation, and
scene understanding. By converting images into rich text descriptions,
nutritionists can assess individual dietary intake based on the captions
instead of the original images, reducing the risk of privacy leakage from
images. To this end, an egocentric dietary image captioning dataset has been
built, which consists of in-the-wild images captured by head-worn and
chest-worn cameras in field studies in Ghana. A novel transformer-based
architecture is designed to caption egocentric dietary images. Comprehensive
experiments have been conducted to evaluate the effectiveness and to justify
the design of the proposed architecture for egocentric dietary image
captioning. To the best of our knowledge, this is the first work that applies
image captioning for dietary intake assessment in real life settings.
| [
{
"version": "v1",
"created": "Thu, 1 Jul 2021 11:16:44 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 08:20:17 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Qiu",
"Jianing",
""
],
[
"Lo",
"Frank P. -W.",
""
],
[
"Gu",
"Xiao",
""
],
[
"Jobarteh",
"Modou L.",
""
],
[
"Jia",
"Wenyan",
""
],
[
"Baranowski",
"Tom",
""
],
[
"Steiner-Asiedu",
"Matilda",
""
],
[
"Anderson",
"Alex K.",
""
],
[
"McCrory",
"Megan A",
""
],
[
"Sazonov",
"Edward",
""
],
[
"Sun",
"Mingui",
""
],
[
"Frost",
"Gary",
""
],
[
"Lo",
"Benny",
""
]
] | TITLE: Egocentric Image Captioning for Privacy-Preserved Passive Dietary Intake
Monitoring
ABSTRACT: Camera-based passive dietary intake monitoring is able to continuously
capture the eating episodes of a subject, recording rich visual information,
such as the type and volume of food being consumed, as well as the eating
behaviours of the subject. However, there currently is no method that is able
to incorporate these visual clues and provide a comprehensive context of
dietary intake from passive recording (e.g., is the subject sharing food with
others, what food the subject is eating, and how much food is left in the
bowl). On the other hand, privacy is a major concern while egocentric wearable
cameras are used for capturing. In this paper, we propose a privacy-preserved
secure solution (i.e., egocentric image captioning) for dietary assessment with
passive monitoring, which unifies food recognition, volume estimation, and
scene understanding. By converting images into rich text descriptions,
nutritionists can assess individual dietary intake based on the captions
instead of the original images, reducing the risk of privacy leakage from
images. To this end, an egocentric dietary image captioning dataset has been
built, which consists of in-the-wild images captured by head-worn and
chest-worn cameras in field studies in Ghana. A novel transformer-based
architecture is designed to caption egocentric dietary images. Comprehensive
experiments have been conducted to evaluate the effectiveness and to justify
the design of the proposed architecture for egocentric dietary image
captioning. To the best of our knowledge, this is the first work that applies
image captioning for dietary intake assessment in real life settings.
|
2107.06707 | Ning Ma | Ning Ma, Jiajun Bu, Zhen Zhang, Sheng Zhou | Uncertainty-Guided Mixup for Semi-Supervised Domain Adaptation without
Source Data | null | Volume 262, 28 February 2023, 110208 | 10.1016/j.knosys.2022.110208 | 11 | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Present domain adaptation methods usually perform explicit representation
alignment by simultaneously accessing the source data and target data. However,
the source data are not always available due to the privacy preserving
consideration or bandwidth limitation. Source-free domain adaptation aims to
solve the above problem by performing domain adaptation without accessing the
source data. The adaptation paradigm is receiving more and more attention in
recent years, and multiple works have been proposed for unsupervised
source-free domain adaptation. However, without utilizing any supervised signal
and source data at the adaptation stage, the optimization of the target model
is unstable and fragile. To alleviate the problem, we focus on semi-supervised
domain adaptation under source-free setting. More specifically, we propose
uncertainty-guided Mixup to reduce the representation's intra-domain
discrepancy and perform inter-domain alignment without directly accessing the
source data. Finally, we conduct extensive semi-supervised domain adaptation
experiments on various datasets. Our method outperforms the recent
semi-supervised baselines and the unsupervised variant also achieves
competitive performance. The experiment codes will be released in the future.
| [
{
"version": "v1",
"created": "Wed, 14 Jul 2021 13:54:02 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Ma",
"Ning",
""
],
[
"Bu",
"Jiajun",
""
],
[
"Zhang",
"Zhen",
""
],
[
"Zhou",
"Sheng",
""
]
] | TITLE: Uncertainty-Guided Mixup for Semi-Supervised Domain Adaptation without
Source Data
ABSTRACT: Present domain adaptation methods usually perform explicit representation
alignment by simultaneously accessing the source data and target data. However,
the source data are not always available due to the privacy preserving
consideration or bandwidth limitation. Source-free domain adaptation aims to
solve the above problem by performing domain adaptation without accessing the
source data. The adaptation paradigm is receiving more and more attention in
recent years, and multiple works have been proposed for unsupervised
source-free domain adaptation. However, without utilizing any supervised signal
and source data at the adaptation stage, the optimization of the target model
is unstable and fragile. To alleviate the problem, we focus on semi-supervised
domain adaptation under source-free setting. More specifically, we propose
uncertainty-guided Mixup to reduce the representation's intra-domain
discrepancy and perform inter-domain alignment without directly accessing the
source data. Finally, we conduct extensive semi-supervised domain adaptation
experiments on various datasets. Our method outperforms the recent
semi-supervised baselines and the unsupervised variant also achieves
competitive performance. The experiment codes will be released in the future.
|
2110.05088 | Hiroshi Sakamoto | Shinji Ono and Jun Takata and Masaharu Kataoka and Tomohiro I and
Kilho Shin and Hiroshi Sakamoto | Privacy-Preserving Feature Selection with Fully Homomorphic Encryption | 14 pages | Algorithms 15(7), Article number 229, 2022 | 10.3390/a15070229 | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For the feature selection problem, we propose an efficient privacy-preserving
algorithm. Let $D$, $F$, and $C$ be data, feature, and class sets,
respectively, where the feature value $x(F_i)$ and the class label $x(C)$ are
given for each $x\in D$ and $F_i \in F$. For a triple $(D,F,C)$, the feature
selection problem is to find a consistent and minimal subset $F' \subseteq F$,
where `consistent' means that, for any $x,y\in D$, $x(C)=y(C)$ if
$x(F_i)=y(F_i)$ for $F_i\in F'$, and `minimal' means that any proper subset of
$F'$ is no longer consistent. On distributed datasets, we consider feature
selection as a privacy-preserving problem: Assume that semi-honest parties
$\textsf A$ and $\textsf B$ have their own personal $D_{\textsf A}$ and
$D_{\textsf B}$. The goal is to solve the feature selection problem for
$D_{\textsf A}\cup D_{\textsf B}$ without revealing their privacy. In this
paper, we propose a secure and efficient algorithm based on fully homomorphic
encryption, and we implement our algorithm to show its effectiveness for
various practical data. The proposed algorithm is the first one that can
directly simulate the CWC (Combination of Weakest Components) algorithm on
ciphertext, which is one of the best performers for the feature selection
problem on the plaintext.
| [
{
"version": "v1",
"created": "Mon, 11 Oct 2021 08:48:32 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Feb 2022 07:40:09 GMT"
},
{
"version": "v3",
"created": "Thu, 2 Jun 2022 03:08:46 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Ono",
"Shinji",
""
],
[
"Takata",
"Jun",
""
],
[
"Kataoka",
"Masaharu",
""
],
[
"I",
"Tomohiro",
""
],
[
"Shin",
"Kilho",
""
],
[
"Sakamoto",
"Hiroshi",
""
]
] | TITLE: Privacy-Preserving Feature Selection with Fully Homomorphic Encryption
ABSTRACT: For the feature selection problem, we propose an efficient privacy-preserving
algorithm. Let $D$, $F$, and $C$ be data, feature, and class sets,
respectively, where the feature value $x(F_i)$ and the class label $x(C)$ are
given for each $x\in D$ and $F_i \in F$. For a triple $(D,F,C)$, the feature
selection problem is to find a consistent and minimal subset $F' \subseteq F$,
where `consistent' means that, for any $x,y\in D$, $x(C)=y(C)$ if
$x(F_i)=y(F_i)$ for $F_i\in F'$, and `minimal' means that any proper subset of
$F'$ is no longer consistent. On distributed datasets, we consider feature
selection as a privacy-preserving problem: Assume that semi-honest parties
$\textsf A$ and $\textsf B$ have their own personal $D_{\textsf A}$ and
$D_{\textsf B}$. The goal is to solve the feature selection problem for
$D_{\textsf A}\cup D_{\textsf B}$ without revealing their privacy. In this
paper, we propose a secure and efficient algorithm based on fully homomorphic
encryption, and we implement our algorithm to show its effectiveness for
various practical data. The proposed algorithm is the first one that can
directly simulate the CWC (Combination of Weakest Components) algorithm on
ciphertext, which is one of the best performers for the feature selection
problem on the plaintext.
|
2110.09231 | Nathan Sanders | Nathan E Sanders, Bruce Schneier | Machine Learning Featurizations for AI Hacking of Political Systems | 11 2-column pages, 2 figures | null | null | null | cs.CY cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | What would the inputs be to a machine whose output is the destabilization of
a robust democracy, or whose emanations could disrupt the political power of
nations? In the recent essay "The Coming AI Hackers," Schneier (2021) proposed
a future application of artificial intelligences to discover, manipulate, and
exploit vulnerabilities of social, economic, and political systems at speeds
far greater than humans' ability to recognize and respond to such threats. This
work advances the concept by applying to it theory from machine learning,
hypothesizing some possible "featurization" (input specification and
transformation) frameworks for AI hacking. Focusing on the political domain, we
develop graph and sequence data representations that would enable the
application of a range of deep learning models to predict attributes and
outcomes of political, particularly legislative, systems. We explore possible
data models, datasets, predictive tasks, and actionable applications associated
with each framework. We speculate about the likely practical impact and
feasibility of such models, and conclude by discussing their ethical
implications.
| [
{
"version": "v1",
"created": "Fri, 8 Oct 2021 16:51:31 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 23:05:22 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Sanders",
"Nathan E",
""
],
[
"Schneier",
"Bruce",
""
]
] | TITLE: Machine Learning Featurizations for AI Hacking of Political Systems
ABSTRACT: What would the inputs be to a machine whose output is the destabilization of
a robust democracy, or whose emanations could disrupt the political power of
nations? In the recent essay "The Coming AI Hackers," Schneier (2021) proposed
a future application of artificial intelligences to discover, manipulate, and
exploit vulnerabilities of social, economic, and political systems at speeds
far greater than humans' ability to recognize and respond to such threats. This
work advances the concept by applying to it theory from machine learning,
hypothesizing some possible "featurization" (input specification and
transformation) frameworks for AI hacking. Focusing on the political domain, we
develop graph and sequence data representations that would enable the
application of a range of deep learning models to predict attributes and
outcomes of political, particularly legislative, systems. We explore possible
data models, datasets, predictive tasks, and actionable applications associated
with each framework. We speculate about the likely practical impact and
feasibility of such models, and conclude by discussing their ethical
implications.
|
2111.01361 | Sloan Nietert | Sloan Nietert, Rachel Cummings, Ziv Goldfeld | Outlier-Robust Optimal Transport: Duality, Structure, and Statistical
Analysis | updated to match AISTATS publication | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Wasserstein distance, rooted in optimal transport (OT) theory, is a
popular discrepancy measure between probability distributions with various
applications to statistics and machine learning. Despite their rich structure
and demonstrated utility, Wasserstein distances are sensitive to outliers in
the considered distributions, which hinders applicability in practice. We
propose a new outlier-robust Wasserstein distance $\mathsf{W}_p^\varepsilon$
which allows for $\varepsilon$ outlier mass to be removed from each
contaminated distribution. Under standard moment assumptions,
$\mathsf{W}_p^\varepsilon$ is shown to achieve strong robust estimation
guarantees under the Huber $\varepsilon$-contamination model. Our formulation
of this robust distance amounts to a highly regular optimization problem that
lends itself better for analysis compared to previously considered frameworks.
Leveraging this, we conduct a thorough theoretical study of
$\mathsf{W}_p^\varepsilon$, encompassing robustness guarantees,
characterization of optimal perturbations, regularity, duality, and statistical
estimation. In particular, by decoupling the optimization variables, we arrive
at a simple dual form for $\mathsf{W}_p^\varepsilon$ that can be implemented
via an elementary modification to standard, duality-based OT solvers. We
illustrate the virtues of our framework via applications to generative modeling
with contaminated datasets.
| [
{
"version": "v1",
"created": "Tue, 2 Nov 2021 04:05:45 GMT"
},
{
"version": "v2",
"created": "Fri, 5 Nov 2021 18:31:46 GMT"
},
{
"version": "v3",
"created": "Sun, 27 Feb 2022 01:02:32 GMT"
},
{
"version": "v4",
"created": "Tue, 28 Feb 2023 19:20:45 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Nietert",
"Sloan",
""
],
[
"Cummings",
"Rachel",
""
],
[
"Goldfeld",
"Ziv",
""
]
] | TITLE: Outlier-Robust Optimal Transport: Duality, Structure, and Statistical
Analysis
ABSTRACT: The Wasserstein distance, rooted in optimal transport (OT) theory, is a
popular discrepancy measure between probability distributions with various
applications to statistics and machine learning. Despite their rich structure
and demonstrated utility, Wasserstein distances are sensitive to outliers in
the considered distributions, which hinders applicability in practice. We
propose a new outlier-robust Wasserstein distance $\mathsf{W}_p^\varepsilon$
which allows for $\varepsilon$ outlier mass to be removed from each
contaminated distribution. Under standard moment assumptions,
$\mathsf{W}_p^\varepsilon$ is shown to achieve strong robust estimation
guarantees under the Huber $\varepsilon$-contamination model. Our formulation
of this robust distance amounts to a highly regular optimization problem that
lends itself better for analysis compared to previously considered frameworks.
Leveraging this, we conduct a thorough theoretical study of
$\mathsf{W}_p^\varepsilon$, encompassing robustness guarantees,
characterization of optimal perturbations, regularity, duality, and statistical
estimation. In particular, by decoupling the optimization variables, we arrive
at a simple dual form for $\mathsf{W}_p^\varepsilon$ that can be implemented
via an elementary modification to standard, duality-based OT solvers. We
illustrate the virtues of our framework via applications to generative modeling
with contaminated datasets.
|
2204.10414 | Biswajit Paria | Abhimanyu Das, Weihao Kong, Biswajit Paria, Rajat Sen | Dirichlet Proportions Model for Hierarchically Coherent Probabilistic
Forecasting | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Probabilistic, hierarchically coherent forecasting is a key problem in many
practical forecasting applications -- the goal is to obtain coherent
probabilistic predictions for a large number of time series arranged in a
pre-specified tree hierarchy. In this paper, we present an end-to-end deep
probabilistic model for hierarchical forecasting that is motivated by a
classical top-down strategy. It jointly learns the distribution of the root
time series, and the (dirichlet) proportions according to which each parent
time-series is split among its children at any point in time. The resulting
forecasts are naturally coherent, and provide probabilistic predictions over
all time series in the hierarchy. We experiment on several public datasets and
demonstrate significant improvements of up to 26% on most datasets compared to
state-of-the-art baselines. Finally, we also provide theoretical justification
for the superiority of our top-down approach compared to the more traditional
bottom-up modeling.
| [
{
"version": "v1",
"created": "Thu, 21 Apr 2022 21:32:28 GMT"
},
{
"version": "v2",
"created": "Wed, 19 Oct 2022 05:03:56 GMT"
},
{
"version": "v3",
"created": "Wed, 1 Mar 2023 05:07:42 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Das",
"Abhimanyu",
""
],
[
"Kong",
"Weihao",
""
],
[
"Paria",
"Biswajit",
""
],
[
"Sen",
"Rajat",
""
]
] | TITLE: Dirichlet Proportions Model for Hierarchically Coherent Probabilistic
Forecasting
ABSTRACT: Probabilistic, hierarchically coherent forecasting is a key problem in many
practical forecasting applications -- the goal is to obtain coherent
probabilistic predictions for a large number of time series arranged in a
pre-specified tree hierarchy. In this paper, we present an end-to-end deep
probabilistic model for hierarchical forecasting that is motivated by a
classical top-down strategy. It jointly learns the distribution of the root
time series, and the (dirichlet) proportions according to which each parent
time-series is split among its children at any point in time. The resulting
forecasts are naturally coherent, and provide probabilistic predictions over
all time series in the hierarchy. We experiment on several public datasets and
demonstrate significant improvements of up to 26% on most datasets compared to
state-of-the-art baselines. Finally, we also provide theoretical justification
for the superiority of our top-down approach compared to the more traditional
bottom-up modeling.
|
2204.12637 | Wenbin Song | Wenbin Song, Di Wu, Weiming Shen and Benoit Boulet | Meta-Learning Based Early Fault Detection for Rolling Bearings via
Few-Shot Anomaly Detection | The results require to be updated | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Early fault detection (EFD) of rolling bearings can recognize slight
deviation of the health states and contribute to the stability of mechanical
systems. In practice, very limited target bearing data are available to conduct
EFD, which makes it hard to adapt to the EFD task of new bearings. To address
this problem, many transfer learning based EFD methods utilize historical data
to learn transferable domain knowledge and conduct early fault detection on new
target bearings. However, most existing methods only consider the distribution
drift across different working conditions but ignore the difference between
bearings under the same working condition, which is called Unit-to-Unit
Variability (UtUV). The setting of EFD with limited target data considering
UtUV can be formulated as a Few-shot Anomaly Detection task. Therefore, this
paper proposes a novel EFD method based on meta-learning considering UtUV. The
proposed method can learn a generic metric based on Relation Network (RN) to
measure the similarity between normal data and the new arrival target bearing
data. Besides, the proposed method utilizes a health state embedding strategy
to decrease false alarms. The performance of proposed method is tested on two
bearing datasets. The results show that the proposed method can detect
incipient faults earlier than the baselines with lower false alarms.
| [
{
"version": "v1",
"created": "Wed, 27 Apr 2022 00:02:35 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 22:57:53 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Song",
"Wenbin",
""
],
[
"Wu",
"Di",
""
],
[
"Shen",
"Weiming",
""
],
[
"Boulet",
"Benoit",
""
]
] | TITLE: Meta-Learning Based Early Fault Detection for Rolling Bearings via
Few-Shot Anomaly Detection
ABSTRACT: Early fault detection (EFD) of rolling bearings can recognize slight
deviation of the health states and contribute to the stability of mechanical
systems. In practice, very limited target bearing data are available to conduct
EFD, which makes it hard to adapt to the EFD task of new bearings. To address
this problem, many transfer learning based EFD methods utilize historical data
to learn transferable domain knowledge and conduct early fault detection on new
target bearings. However, most existing methods only consider the distribution
drift across different working conditions but ignore the difference between
bearings under the same working condition, which is called Unit-to-Unit
Variability (UtUV). The setting of EFD with limited target data considering
UtUV can be formulated as a Few-shot Anomaly Detection task. Therefore, this
paper proposes a novel EFD method based on meta-learning considering UtUV. The
proposed method can learn a generic metric based on Relation Network (RN) to
measure the similarity between normal data and the new arrival target bearing
data. Besides, the proposed method utilizes a health state embedding strategy
to decrease false alarms. The performance of proposed method is tested on two
bearing datasets. The results show that the proposed method can detect
incipient faults earlier than the baselines with lower false alarms.
|
2205.11027 | Jianxiong Li | Jianxiong Li, Xianyuan Zhan, Haoran Xu, Xiangyu Zhu, Jingjing Liu,
Ya-Qin Zhang | When Data Geometry Meets Deep Function: Generalizing Offline
Reinforcement Learning | ICLR2023 | null | null | null | cs.LG cs.AI cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In offline reinforcement learning (RL), one detrimental issue to policy
learning is the error accumulation of deep Q function in out-of-distribution
(OOD) areas. Unfortunately, existing offline RL methods are often
over-conservative, inevitably hurting generalization performance outside data
distribution. In our study, one interesting observation is that deep Q
functions approximate well inside the convex hull of training data. Inspired by
this, we propose a new method, DOGE (Distance-sensitive Offline RL with better
GEneralization). DOGE marries dataset geometry with deep function approximators
in offline RL, and enables exploitation in generalizable OOD areas rather than
strictly constraining policy within data distribution. Specifically, DOGE
trains a state-conditioned distance function that can be readily plugged into
standard actor-critic methods as a policy constraint. Simple yet elegant, our
algorithm enjoys better generalization compared to state-of-the-art methods on
D4RL benchmarks. Theoretical analysis demonstrates the superiority of our
approach to existing methods that are solely based on data distribution or
support constraints.
| [
{
"version": "v1",
"created": "Mon, 23 May 2022 04:01:11 GMT"
},
{
"version": "v2",
"created": "Fri, 3 Feb 2023 12:52:38 GMT"
},
{
"version": "v3",
"created": "Wed, 8 Feb 2023 09:12:47 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Li",
"Jianxiong",
""
],
[
"Zhan",
"Xianyuan",
""
],
[
"Xu",
"Haoran",
""
],
[
"Zhu",
"Xiangyu",
""
],
[
"Liu",
"Jingjing",
""
],
[
"Zhang",
"Ya-Qin",
""
]
] | TITLE: When Data Geometry Meets Deep Function: Generalizing Offline
Reinforcement Learning
ABSTRACT: In offline reinforcement learning (RL), one detrimental issue to policy
learning is the error accumulation of deep Q function in out-of-distribution
(OOD) areas. Unfortunately, existing offline RL methods are often
over-conservative, inevitably hurting generalization performance outside data
distribution. In our study, one interesting observation is that deep Q
functions approximate well inside the convex hull of training data. Inspired by
this, we propose a new method, DOGE (Distance-sensitive Offline RL with better
GEneralization). DOGE marries dataset geometry with deep function approximators
in offline RL, and enables exploitation in generalizable OOD areas rather than
strictly constraining policy within data distribution. Specifically, DOGE
trains a state-conditioned distance function that can be readily plugged into
standard actor-critic methods as a policy constraint. Simple yet elegant, our
algorithm enjoys better generalization compared to state-of-the-art methods on
D4RL benchmarks. Theoretical analysis demonstrates the superiority of our
approach to existing methods that are solely based on data distribution or
support constraints.
|
2205.15512 | Han Zhong | Wei Xiong, Han Zhong, Chengshuai Shi, Cong Shen, Liwei Wang, Tong
Zhang | Nearly Minimax Optimal Offline Reinforcement Learning with Linear
Function Approximation: Single-Agent MDP and Markov Game | null | null | null | null | cs.LG cs.GT stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Offline reinforcement learning (RL) aims at learning an optimal strategy
using a pre-collected dataset without further interactions with the
environment. While various algorithms have been proposed for offline RL in the
previous literature, the minimax optimality has only been (nearly) established
for tabular Markov decision processes (MDPs). In this paper, we focus on
offline RL with linear function approximation and propose a new pessimism-based
algorithm for offline linear MDP. At the core of our algorithm is the
uncertainty decomposition via a reference function, which is new in the
literature of offline RL under linear function approximation. Theoretical
analysis demonstrates that our algorithm can match the performance lower bound
up to logarithmic factors. We also extend our techniques to the two-player
zero-sum Markov games (MGs), and establish a new performance lower bound for
MGs, which tightens the existing result, and verifies the nearly minimax
optimality of the proposed algorithm. To the best of our knowledge, these are
the first computationally efficient and nearly minimax optimal algorithms for
offline single-agent MDPs and MGs with linear function approximation.
| [
{
"version": "v1",
"created": "Tue, 31 May 2022 02:50:17 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 13:37:28 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Xiong",
"Wei",
""
],
[
"Zhong",
"Han",
""
],
[
"Shi",
"Chengshuai",
""
],
[
"Shen",
"Cong",
""
],
[
"Wang",
"Liwei",
""
],
[
"Zhang",
"Tong",
""
]
] | TITLE: Nearly Minimax Optimal Offline Reinforcement Learning with Linear
Function Approximation: Single-Agent MDP and Markov Game
ABSTRACT: Offline reinforcement learning (RL) aims at learning an optimal strategy
using a pre-collected dataset without further interactions with the
environment. While various algorithms have been proposed for offline RL in the
previous literature, the minimax optimality has only been (nearly) established
for tabular Markov decision processes (MDPs). In this paper, we focus on
offline RL with linear function approximation and propose a new pessimism-based
algorithm for offline linear MDP. At the core of our algorithm is the
uncertainty decomposition via a reference function, which is new in the
literature of offline RL under linear function approximation. Theoretical
analysis demonstrates that our algorithm can match the performance lower bound
up to logarithmic factors. We also extend our techniques to the two-player
zero-sum Markov games (MGs), and establish a new performance lower bound for
MGs, which tightens the existing result, and verifies the nearly minimax
optimality of the proposed algorithm. To the best of our knowledge, these are
the first computationally efficient and nearly minimax optimal algorithms for
offline single-agent MDPs and MGs with linear function approximation.
|
2206.01729 | Bowen Jing | Bowen Jing, Gabriele Corso, Jeffrey Chang, Regina Barzilay, Tommi
Jaakkola | Torsional Diffusion for Molecular Conformer Generation | NeurIPS 2022 | null | null | null | physics.chem-ph cs.LG q-bio.BM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Molecular conformer generation is a fundamental task in computational
chemistry. Several machine learning approaches have been developed, but none
have outperformed state-of-the-art cheminformatics methods. We propose
torsional diffusion, a novel diffusion framework that operates on the space of
torsion angles via a diffusion process on the hypertorus and an
extrinsic-to-intrinsic score model. On a standard benchmark of drug-like
molecules, torsional diffusion generates superior conformer ensembles compared
to machine learning and cheminformatics methods in terms of both RMSD and
chemical properties, and is orders of magnitude faster than previous
diffusion-based models. Moreover, our model provides exact likelihoods, which
we employ to build the first generalizable Boltzmann generator. Code is
available at https://github.com/gcorso/torsional-diffusion.
| [
{
"version": "v1",
"created": "Wed, 1 Jun 2022 04:30:41 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 02:27:10 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Jing",
"Bowen",
""
],
[
"Corso",
"Gabriele",
""
],
[
"Chang",
"Jeffrey",
""
],
[
"Barzilay",
"Regina",
""
],
[
"Jaakkola",
"Tommi",
""
]
] | TITLE: Torsional Diffusion for Molecular Conformer Generation
ABSTRACT: Molecular conformer generation is a fundamental task in computational
chemistry. Several machine learning approaches have been developed, but none
have outperformed state-of-the-art cheminformatics methods. We propose
torsional diffusion, a novel diffusion framework that operates on the space of
torsion angles via a diffusion process on the hypertorus and an
extrinsic-to-intrinsic score model. On a standard benchmark of drug-like
molecules, torsional diffusion generates superior conformer ensembles compared
to machine learning and cheminformatics methods in terms of both RMSD and
chemical properties, and is orders of magnitude faster than previous
diffusion-based models. Moreover, our model provides exact likelihoods, which
we employ to build the first generalizable Boltzmann generator. Code is
available at https://github.com/gcorso/torsional-diffusion.
|
2206.14507 | Siheon Park | Siheon Park, Daniel K. Park, June-Koo Kevin Rhee | Variational Quantum Approximate Support Vector Machine with Inference
Transfer | 16 pages, 4 figures | Sci Rep 13, 3288 (2023) | 10.1038/s41598-023-29495-y | null | quant-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A kernel-based quantum classifier is the most practical and influential
quantum machine learning technique for the hyper-linear classification of
complex data. We propose a Variational Quantum Approximate Support Vector
Machine (VQASVM) algorithm that demonstrates empirical sub-quadratic run-time
complexity with quantum operations feasible even in NISQ computers. We
experimented our algorithm with toy example dataset on cloud-based NISQ
machines as a proof of concept. We also numerically investigated its
performance on the standard Iris flower and MNIST datasets to confirm the
practicality and scalability.
| [
{
"version": "v1",
"created": "Wed, 29 Jun 2022 09:56:59 GMT"
},
{
"version": "v2",
"created": "Fri, 12 Aug 2022 17:24:02 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Feb 2023 15:53:18 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Park",
"Siheon",
""
],
[
"Park",
"Daniel K.",
""
],
[
"Rhee",
"June-Koo Kevin",
""
]
] | TITLE: Variational Quantum Approximate Support Vector Machine with Inference
Transfer
ABSTRACT: A kernel-based quantum classifier is the most practical and influential
quantum machine learning technique for the hyper-linear classification of
complex data. We propose a Variational Quantum Approximate Support Vector
Machine (VQASVM) algorithm that demonstrates empirical sub-quadratic run-time
complexity with quantum operations feasible even in NISQ computers. We
experimented our algorithm with toy example dataset on cloud-based NISQ
machines as a proof of concept. We also numerically investigated its
performance on the standard Iris flower and MNIST datasets to confirm the
practicality and scalability.
|
2206.15119 | Alberto Bertipaglia | A. Bertipaglia (1), D. de Mol (1), M. Alirezaei (2 and 3), R. Happee
(1), B. Shyrokau (1) ((1) Delft University of Technology, (2) Eindhoven
University of Technology, (3) Siemens Digital Industries Software) | Model-based vs Data-driven Estimation of Vehicle Sideslip Angle and
Benefits of Tyre Force Measurements | Published in the 15th International Symposium on Advanced Vehicle
Control, Kanagawa, Japan, 2022 | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper provides a comprehensive comparison of model-based and data-driven
approaches and analyses the benefits of using measured tyre forces for vehicle
sideslip angle estimation. The model-based approaches are based on an extended
Kalman filter and an unscented Kalman filter, in which the measured tyre forces
are utilised in the observation model. An adaptive covariance matrix is
introduced to minimise the tyre model mismatch during evasive manoeuvres. For
data-driven approaches, feed forward and recurrent neural networks are
evaluated. Both approaches use the standard inertial measurement unit and the
tyre force measurements as inputs. Using the large-scale experimental dataset
of 216 manoeuvres, we demonstrate a significant improvement in accuracy using
data-driven vs. model-based approaches. Tyre force measurements improve the
performance of both model-based and data-driven approaches, especially in the
non-linear regime of tyres.
| [
{
"version": "v1",
"created": "Thu, 30 Jun 2022 08:30:12 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 14:04:07 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Bertipaglia",
"A.",
"",
"2 and 3"
],
[
"de Mol",
"D.",
"",
"2 and 3"
],
[
"Alirezaei",
"M.",
"",
"2 and 3"
],
[
"Happee",
"R.",
""
],
[
"Shyrokau",
"B.",
""
]
] | TITLE: Model-based vs Data-driven Estimation of Vehicle Sideslip Angle and
Benefits of Tyre Force Measurements
ABSTRACT: This paper provides a comprehensive comparison of model-based and data-driven
approaches and analyses the benefits of using measured tyre forces for vehicle
sideslip angle estimation. The model-based approaches are based on an extended
Kalman filter and an unscented Kalman filter, in which the measured tyre forces
are utilised in the observation model. An adaptive covariance matrix is
introduced to minimise the tyre model mismatch during evasive manoeuvres. For
data-driven approaches, feed forward and recurrent neural networks are
evaluated. Both approaches use the standard inertial measurement unit and the
tyre force measurements as inputs. Using the large-scale experimental dataset
of 216 manoeuvres, we demonstrate a significant improvement in accuracy using
data-driven vs. model-based approaches. Tyre force measurements improve the
performance of both model-based and data-driven approaches, especially in the
non-linear regime of tyres.
|
2207.05833 | Xingjian Shi | Zhihan Gao, Xingjian Shi, Hao Wang, Yi Zhu, Yuyang Wang, Mu Li,
Dit-Yan Yeung | Earthformer: Exploring Space-Time Transformers for Earth System
Forecasting | Published at NeurIPS 2022. Camera-ready version | null | null | null | cs.LG cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conventionally, Earth system (e.g., weather and climate) forecasting relies
on numerical simulation with complex physical models and are hence both
expensive in computation and demanding on domain expertise. With the explosive
growth of the spatiotemporal Earth observation data in the past decade,
data-driven models that apply Deep Learning (DL) are demonstrating impressive
potential for various Earth system forecasting tasks. The Transformer as an
emerging DL architecture, despite its broad success in other domains, has
limited adoption in this area. In this paper, we propose Earthformer, a
space-time Transformer for Earth system forecasting. Earthformer is based on a
generic, flexible and efficient space-time attention block, named Cuboid
Attention. The idea is to decompose the data into cuboids and apply
cuboid-level self-attention in parallel. These cuboids are further connected
with a collection of global vectors. We conduct experiments on the MovingMNIST
dataset and a newly proposed chaotic N-body MNIST dataset to verify the
effectiveness of cuboid attention and figure out the best design of
Earthformer. Experiments on two real-world benchmarks about precipitation
nowcasting and El Nino/Southern Oscillation (ENSO) forecasting show Earthformer
achieves state-of-the-art performance. Code is available:
https://github.com/amazon-science/earth-forecasting-transformer .
| [
{
"version": "v1",
"created": "Tue, 12 Jul 2022 20:52:26 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 04:40:21 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Gao",
"Zhihan",
""
],
[
"Shi",
"Xingjian",
""
],
[
"Wang",
"Hao",
""
],
[
"Zhu",
"Yi",
""
],
[
"Wang",
"Yuyang",
""
],
[
"Li",
"Mu",
""
],
[
"Yeung",
"Dit-Yan",
""
]
] | TITLE: Earthformer: Exploring Space-Time Transformers for Earth System
Forecasting
ABSTRACT: Conventionally, Earth system (e.g., weather and climate) forecasting relies
on numerical simulation with complex physical models and are hence both
expensive in computation and demanding on domain expertise. With the explosive
growth of the spatiotemporal Earth observation data in the past decade,
data-driven models that apply Deep Learning (DL) are demonstrating impressive
potential for various Earth system forecasting tasks. The Transformer as an
emerging DL architecture, despite its broad success in other domains, has
limited adoption in this area. In this paper, we propose Earthformer, a
space-time Transformer for Earth system forecasting. Earthformer is based on a
generic, flexible and efficient space-time attention block, named Cuboid
Attention. The idea is to decompose the data into cuboids and apply
cuboid-level self-attention in parallel. These cuboids are further connected
with a collection of global vectors. We conduct experiments on the MovingMNIST
dataset and a newly proposed chaotic N-body MNIST dataset to verify the
effectiveness of cuboid attention and figure out the best design of
Earthformer. Experiments on two real-world benchmarks about precipitation
nowcasting and El Nino/Southern Oscillation (ENSO) forecasting show Earthformer
achieves state-of-the-art performance. Code is available:
https://github.com/amazon-science/earth-forecasting-transformer .
|
2207.09217 | Yinghui Li | Ding Zhang, Yinghui Li, Qingyu Zhou, Shirong Ma, Yangning Li, Yunbo
Cao, Hai-Tao Zheng | Contextual Similarity is More Valuable than Character Similarity: An
Empirical Study for Chinese Spell Checking | Accepted by ICASSP2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Chinese Spell Checking (CSC) task aims to detect and correct Chinese spelling
errors. Recently, related researches focus on introducing character similarity
from confusion set to enhance the CSC models, ignoring the context of
characters that contain richer information. To make better use of contextual
information, we propose a simple yet effective Curriculum Learning (CL)
framework for the CSC task. With the help of our model-agnostic CL framework,
existing CSC models will be trained from easy to difficult as humans learn
Chinese characters and achieve further performance improvements. Extensive
experiments and detailed analyses on widely used SIGHAN datasets show that our
method outperforms previous state-of-the-art methods. More instructively, our
study empirically suggests that contextual similarity is more valuable than
character similarity for the CSC task.
| [
{
"version": "v1",
"created": "Sun, 17 Jul 2022 03:12:27 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Oct 2022 07:21:32 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Feb 2023 09:47:49 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Zhang",
"Ding",
""
],
[
"Li",
"Yinghui",
""
],
[
"Zhou",
"Qingyu",
""
],
[
"Ma",
"Shirong",
""
],
[
"Li",
"Yangning",
""
],
[
"Cao",
"Yunbo",
""
],
[
"Zheng",
"Hai-Tao",
""
]
] | TITLE: Contextual Similarity is More Valuable than Character Similarity: An
Empirical Study for Chinese Spell Checking
ABSTRACT: Chinese Spell Checking (CSC) task aims to detect and correct Chinese spelling
errors. Recently, related researches focus on introducing character similarity
from confusion set to enhance the CSC models, ignoring the context of
characters that contain richer information. To make better use of contextual
information, we propose a simple yet effective Curriculum Learning (CL)
framework for the CSC task. With the help of our model-agnostic CL framework,
existing CSC models will be trained from easy to difficult as humans learn
Chinese characters and achieve further performance improvements. Extensive
experiments and detailed analyses on widely used SIGHAN datasets show that our
method outperforms previous state-of-the-art methods. More instructively, our
study empirically suggests that contextual similarity is more valuable than
character similarity for the CSC task.
|
2207.12988 | Tai Wang | Tai Wang, Jiangmiao Pang, Dahua Lin | Monocular 3D Object Detection with Depth from Motion | ECCV 2022 Oral | null | null | null | cs.CV cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Perceiving 3D objects from monocular inputs is crucial for robotic systems,
given its economy compared to multi-sensor settings. It is notably difficult as
a single image can not provide any clues for predicting absolute depth values.
Motivated by binocular methods for 3D object detection, we take advantage of
the strong geometry structure provided by camera ego-motion for accurate object
depth estimation and detection. We first make a theoretical analysis on this
general two-view case and notice two challenges: 1) Cumulative errors from
multiple estimations that make the direct prediction intractable; 2) Inherent
dilemmas caused by static cameras and matching ambiguity. Accordingly, we
establish the stereo correspondence with a geometry-aware cost volume as the
alternative for depth estimation and further compensate it with monocular
understanding to address the second problem. Our framework, named Depth from
Motion (DfM), then uses the established geometry to lift 2D image features to
the 3D space and detects 3D objects thereon. We also present a pose-free DfM to
make it usable when the camera pose is unavailable. Our framework outperforms
state-of-the-art methods by a large margin on the KITTI benchmark. Detailed
quantitative and qualitative analyses also validate our theoretical
conclusions. The code will be released at
https://github.com/Tai-Wang/Depth-from-Motion.
| [
{
"version": "v1",
"created": "Tue, 26 Jul 2022 15:48:46 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 07:41:59 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Wang",
"Tai",
""
],
[
"Pang",
"Jiangmiao",
""
],
[
"Lin",
"Dahua",
""
]
] | TITLE: Monocular 3D Object Detection with Depth from Motion
ABSTRACT: Perceiving 3D objects from monocular inputs is crucial for robotic systems,
given its economy compared to multi-sensor settings. It is notably difficult as
a single image can not provide any clues for predicting absolute depth values.
Motivated by binocular methods for 3D object detection, we take advantage of
the strong geometry structure provided by camera ego-motion for accurate object
depth estimation and detection. We first make a theoretical analysis on this
general two-view case and notice two challenges: 1) Cumulative errors from
multiple estimations that make the direct prediction intractable; 2) Inherent
dilemmas caused by static cameras and matching ambiguity. Accordingly, we
establish the stereo correspondence with a geometry-aware cost volume as the
alternative for depth estimation and further compensate it with monocular
understanding to address the second problem. Our framework, named Depth from
Motion (DfM), then uses the established geometry to lift 2D image features to
the 3D space and detects 3D objects thereon. We also present a pose-free DfM to
make it usable when the camera pose is unavailable. Our framework outperforms
state-of-the-art methods by a large margin on the KITTI benchmark. Detailed
quantitative and qualitative analyses also validate our theoretical
conclusions. The code will be released at
https://github.com/Tai-Wang/Depth-from-Motion.
|
2208.04202 | Ting Chen | Ting Chen, Ruixiang Zhang, Geoffrey Hinton | Analog Bits: Generating Discrete Data using Diffusion Models with
Self-Conditioning | ICLR'23 | null | null | null | cs.CV cs.AI cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present Bit Diffusion: a simple and generic approach for generating
discrete data with continuous state and continuous time diffusion models. The
main idea behind our approach is to first represent the discrete data as binary
bits, and then train a continuous diffusion model to model these bits as real
numbers which we call analog bits. To generate samples, the model first
generates the analog bits, which are then thresholded to obtain the bits that
represent the discrete variables. We further propose two simple techniques,
namely Self-Conditioning and Asymmetric Time Intervals, which lead to a
significant improvement in sample quality. Despite its simplicity, the proposed
approach can achieve strong performance in both discrete image generation and
image captioning tasks. For discrete image generation, we significantly improve
previous state-of-the-art on both CIFAR-10 (which has 3K discrete 8-bit tokens)
and ImageNet-64x64 (which has 12K discrete 8-bit tokens), outperforming the
best autoregressive model in both sample quality (measured by FID) and
efficiency. For image captioning on MS-COCO dataset, our approach achieves
competitive results compared to autoregressive models.
| [
{
"version": "v1",
"created": "Mon, 8 Aug 2022 15:08:40 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 00:18:33 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Chen",
"Ting",
""
],
[
"Zhang",
"Ruixiang",
""
],
[
"Hinton",
"Geoffrey",
""
]
] | TITLE: Analog Bits: Generating Discrete Data using Diffusion Models with
Self-Conditioning
ABSTRACT: We present Bit Diffusion: a simple and generic approach for generating
discrete data with continuous state and continuous time diffusion models. The
main idea behind our approach is to first represent the discrete data as binary
bits, and then train a continuous diffusion model to model these bits as real
numbers which we call analog bits. To generate samples, the model first
generates the analog bits, which are then thresholded to obtain the bits that
represent the discrete variables. We further propose two simple techniques,
namely Self-Conditioning and Asymmetric Time Intervals, which lead to a
significant improvement in sample quality. Despite its simplicity, the proposed
approach can achieve strong performance in both discrete image generation and
image captioning tasks. For discrete image generation, we significantly improve
previous state-of-the-art on both CIFAR-10 (which has 3K discrete 8-bit tokens)
and ImageNet-64x64 (which has 12K discrete 8-bit tokens), outperforming the
best autoregressive model in both sample quality (measured by FID) and
efficiency. For image captioning on MS-COCO dataset, our approach achieves
competitive results compared to autoregressive models.
|
2208.06521 | Daniel Chui | Daniel Chui, Jason Hartline, James R. Wright | Non-strategic Econometrics (for Initial Play) | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modelling agent preferences has applications in a range of fields including
economics and increasingly, artificial intelligence. These preferences are not
always known and thus may need to be estimated from observed behavior, in which
case a model is required to map agent preferences to behavior, also known as
structural estimation. Traditional models are based on the assumption that
agents are perfectly rational: that is, they perfectly optimize and behave in
accordance with their own interests. Work in the field of behavioral game
theory has shown, however, that human agents often make decisions that are
imperfectly rational, and the field has developed models that relax the perfect
rationality assumption. We apply models developed for predicting behavior
towards estimating preferences and show that they outperform both traditional
and commonly used benchmark models on data collected from human subjects. In
fact, Nash equilibrium and its relaxation, quantal response equilibrium (QRE),
can induce an inaccurate estimate of agent preferences when compared against
ground truth.
A key finding is that modelling non-strategic behavior, conventionally
considered uniform noise, is important for estimating preferences. To this end,
we introduce quantal-linear4, a rich non-strategic model. We also propose an
augmentation to the popular quantal response equilibrium with a non-strategic
component. We call this augmented model QRE+L0 and find an improvement in
estimating values over the standard QRE.
QRE+L0 allows for alternative models of non-strategic behavior in addition to
quantal-linear4.
| [
{
"version": "v1",
"created": "Fri, 12 Aug 2022 22:37:10 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 03:55:23 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Chui",
"Daniel",
""
],
[
"Hartline",
"Jason",
""
],
[
"Wright",
"James R.",
""
]
] | TITLE: Non-strategic Econometrics (for Initial Play)
ABSTRACT: Modelling agent preferences has applications in a range of fields including
economics and increasingly, artificial intelligence. These preferences are not
always known and thus may need to be estimated from observed behavior, in which
case a model is required to map agent preferences to behavior, also known as
structural estimation. Traditional models are based on the assumption that
agents are perfectly rational: that is, they perfectly optimize and behave in
accordance with their own interests. Work in the field of behavioral game
theory has shown, however, that human agents often make decisions that are
imperfectly rational, and the field has developed models that relax the perfect
rationality assumption. We apply models developed for predicting behavior
towards estimating preferences and show that they outperform both traditional
and commonly used benchmark models on data collected from human subjects. In
fact, Nash equilibrium and its relaxation, quantal response equilibrium (QRE),
can induce an inaccurate estimate of agent preferences when compared against
ground truth.
A key finding is that modelling non-strategic behavior, conventionally
considered uniform noise, is important for estimating preferences. To this end,
we introduce quantal-linear4, a rich non-strategic model. We also propose an
augmentation to the popular quantal response equilibrium with a non-strategic
component. We call this augmented model QRE+L0 and find an improvement in
estimating values over the standard QRE.
QRE+L0 allows for alternative models of non-strategic behavior in addition to
quantal-linear4.
|
2208.12428 | Prashant Pandey | Prashant Pandey, Mustafa Chasmai, Tanuj Sur, Brejesh Lall | Robust Prototypical Few-Shot Organ Segmentation with Regularized
Neural-ODEs | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Despite the tremendous progress made by deep learning models in image
semantic segmentation, they typically require large annotated examples, and
increasing attention is being diverted to problem settings like Few-Shot
Learning (FSL) where only a small amount of annotation is needed for
generalisation to novel classes. This is especially seen in medical domains
where dense pixel-level annotations are expensive to obtain. In this paper, we
propose Regularized Prototypical Neural Ordinary Differential Equation
(R-PNODE), a method that leverages intrinsic properties of Neural-ODEs,
assisted and enhanced by additional cluster and consistency losses to perform
Few-Shot Segmentation (FSS) of organs. R-PNODE constrains support and query
features from the same classes to lie closer in the representation space
thereby improving the performance over the existing Convolutional Neural
Network (CNN) based FSS methods. We further demonstrate that while many
existing Deep CNN based methods tend to be extremely vulnerable to adversarial
attacks, R-PNODE exhibits increased adversarial robustness for a wide array of
these attacks. We experiment with three publicly available multi-organ
segmentation datasets in both in-domain and cross-domain FSS settings to
demonstrate the efficacy of our method. In addition, we perform experiments
with seven commonly used adversarial attacks in various settings to demonstrate
R-PNODE's robustness. R-PNODE outperforms the baselines for FSS by significant
margins and also shows superior performance for a wide array of attacks varying
in intensity and design.
| [
{
"version": "v1",
"created": "Fri, 26 Aug 2022 03:53:04 GMT"
},
{
"version": "v2",
"created": "Sat, 17 Sep 2022 04:38:45 GMT"
},
{
"version": "v3",
"created": "Wed, 1 Mar 2023 08:05:18 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Pandey",
"Prashant",
""
],
[
"Chasmai",
"Mustafa",
""
],
[
"Sur",
"Tanuj",
""
],
[
"Lall",
"Brejesh",
""
]
] | TITLE: Robust Prototypical Few-Shot Organ Segmentation with Regularized
Neural-ODEs
ABSTRACT: Despite the tremendous progress made by deep learning models in image
semantic segmentation, they typically require large annotated examples, and
increasing attention is being diverted to problem settings like Few-Shot
Learning (FSL) where only a small amount of annotation is needed for
generalisation to novel classes. This is especially seen in medical domains
where dense pixel-level annotations are expensive to obtain. In this paper, we
propose Regularized Prototypical Neural Ordinary Differential Equation
(R-PNODE), a method that leverages intrinsic properties of Neural-ODEs,
assisted and enhanced by additional cluster and consistency losses to perform
Few-Shot Segmentation (FSS) of organs. R-PNODE constrains support and query
features from the same classes to lie closer in the representation space
thereby improving the performance over the existing Convolutional Neural
Network (CNN) based FSS methods. We further demonstrate that while many
existing Deep CNN based methods tend to be extremely vulnerable to adversarial
attacks, R-PNODE exhibits increased adversarial robustness for a wide array of
these attacks. We experiment with three publicly available multi-organ
segmentation datasets in both in-domain and cross-domain FSS settings to
demonstrate the efficacy of our method. In addition, we perform experiments
with seven commonly used adversarial attacks in various settings to demonstrate
R-PNODE's robustness. R-PNODE outperforms the baselines for FSS by significant
margins and also shows superior performance for a wide array of attacks varying
in intensity and design.
|
2209.00588 | Eloi Alonso | Vincent Micheli, Eloi Alonso, Fran\c{c}ois Fleuret | Transformers are Sample-Efficient World Models | ICLR 2023 (notable top 5%) | null | null | null | cs.LG cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep reinforcement learning agents are notoriously sample inefficient, which
considerably limits their application to real-world problems. Recently, many
model-based methods have been designed to address this issue, with learning in
the imagination of a world model being one of the most prominent approaches.
However, while virtually unlimited interaction with a simulated environment
sounds appealing, the world model has to be accurate over extended periods of
time. Motivated by the success of Transformers in sequence modeling tasks, we
introduce IRIS, a data-efficient agent that learns in a world model composed of
a discrete autoencoder and an autoregressive Transformer. With the equivalent
of only two hours of gameplay in the Atari 100k benchmark, IRIS achieves a mean
human normalized score of 1.046, and outperforms humans on 10 out of 26 games,
setting a new state of the art for methods without lookahead search. To foster
future research on Transformers and world models for sample-efficient
reinforcement learning, we release our code and models at
https://github.com/eloialonso/iris.
| [
{
"version": "v1",
"created": "Thu, 1 Sep 2022 17:03:07 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 09:21:14 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Micheli",
"Vincent",
""
],
[
"Alonso",
"Eloi",
""
],
[
"Fleuret",
"François",
""
]
] | TITLE: Transformers are Sample-Efficient World Models
ABSTRACT: Deep reinforcement learning agents are notoriously sample inefficient, which
considerably limits their application to real-world problems. Recently, many
model-based methods have been designed to address this issue, with learning in
the imagination of a world model being one of the most prominent approaches.
However, while virtually unlimited interaction with a simulated environment
sounds appealing, the world model has to be accurate over extended periods of
time. Motivated by the success of Transformers in sequence modeling tasks, we
introduce IRIS, a data-efficient agent that learns in a world model composed of
a discrete autoencoder and an autoregressive Transformer. With the equivalent
of only two hours of gameplay in the Atari 100k benchmark, IRIS achieves a mean
human normalized score of 1.046, and outperforms humans on 10 out of 26 games,
setting a new state of the art for methods without lookahead search. To foster
future research on Transformers and world models for sample-efficient
reinforcement learning, we release our code and models at
https://github.com/eloialonso/iris.
|
2209.06316 | Islam Ali | Islam Ali, Hong Zhang | Optimizing SLAM Evaluation Footprint Through Dynamic Range Coverage
Analysis of Datasets | null | null | null | null | cs.RO cs.CV | http://creativecommons.org/licenses/by/4.0/ | Simultaneous Localization and Mapping (SLAM) is considered an ever-evolving
problem due to its usage in many applications. Evaluation of SLAM is done
typically using publicly available datasets which are increasing in number and
the level of difficulty. Each dataset provides a certain level of dynamic range
coverage that is a key aspect of measuring the robustness and resilience of
SLAM. In this paper, we provide a systematic analysis of the dynamic range
coverage of datasets based on a number of characterization metrics, and our
analysis shows a huge level of redundancy within and between datasets.
Subsequently, we propose a dynamic programming (DP) algorithm for eliminating
the redundancy in the evaluation process of SLAM by selecting a subset of
sequences that matches a single or multiple dynamic range coverage objectives.
It is shown that, with the help of dataset characterization and DP selection
algorithm, a reduction in the evaluation effort can be achieved while
maintaining the same level of coverage. We also study how the evaluation
process of a real-world SLAM system can be optimized utilizing the method
proposed.
| [
{
"version": "v1",
"created": "Tue, 13 Sep 2022 21:47:18 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 16:10:00 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Ali",
"Islam",
""
],
[
"Zhang",
"Hong",
""
]
] | TITLE: Optimizing SLAM Evaluation Footprint Through Dynamic Range Coverage
Analysis of Datasets
ABSTRACT: Simultaneous Localization and Mapping (SLAM) is considered an ever-evolving
problem due to its usage in many applications. Evaluation of SLAM is done
typically using publicly available datasets which are increasing in number and
the level of difficulty. Each dataset provides a certain level of dynamic range
coverage that is a key aspect of measuring the robustness and resilience of
SLAM. In this paper, we provide a systematic analysis of the dynamic range
coverage of datasets based on a number of characterization metrics, and our
analysis shows a huge level of redundancy within and between datasets.
Subsequently, we propose a dynamic programming (DP) algorithm for eliminating
the redundancy in the evaluation process of SLAM by selecting a subset of
sequences that matches a single or multiple dynamic range coverage objectives.
It is shown that, with the help of dataset characterization and DP selection
algorithm, a reduction in the evaluation effort can be achieved while
maintaining the same level of coverage. We also study how the evaluation
process of a real-world SLAM system can be optimized utilizing the method
proposed.
|
2209.07529 | Haeyong Kang | Haeyong Kang, Jaehong Yoon, Sultan Rizky Hikmawan Madjid, Sung Ju
Hwang, Chang D. Yoo | On the Soft-Subnetwork for Few-shot Class Incremental Learning | The Eleventh International Conference on Learning Representations
(ICLR, 2023) | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Inspired by Regularized Lottery Ticket Hypothesis (RLTH), which hypothesizes
that there exist smooth (non-binary) subnetworks within a dense network that
achieve the competitive performance of the dense network, we propose a few-shot
class incremental learning (FSCIL) method referred to as \emph{Soft-SubNetworks
(SoftNet)}. Our objective is to learn a sequence of sessions incrementally,
where each session only includes a few training instances per class while
preserving the knowledge of the previously learned ones. SoftNet jointly learns
the model weights and adaptive non-binary soft masks at a base training session
in which each mask consists of the major and minor subnetwork; the former aims
to minimize catastrophic forgetting during training, and the latter aims to
avoid overfitting to a few samples in each new training session. We provide
comprehensive empirical validations demonstrating that our SoftNet effectively
tackles the few-shot incremental learning problem by surpassing the performance
of state-of-the-art baselines over benchmark datasets.
| [
{
"version": "v1",
"created": "Thu, 15 Sep 2022 04:54:02 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 12:21:06 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Kang",
"Haeyong",
""
],
[
"Yoon",
"Jaehong",
""
],
[
"Madjid",
"Sultan Rizky Hikmawan",
""
],
[
"Hwang",
"Sung Ju",
""
],
[
"Yoo",
"Chang D.",
""
]
] | TITLE: On the Soft-Subnetwork for Few-shot Class Incremental Learning
ABSTRACT: Inspired by Regularized Lottery Ticket Hypothesis (RLTH), which hypothesizes
that there exist smooth (non-binary) subnetworks within a dense network that
achieve the competitive performance of the dense network, we propose a few-shot
class incremental learning (FSCIL) method referred to as \emph{Soft-SubNetworks
(SoftNet)}. Our objective is to learn a sequence of sessions incrementally,
where each session only includes a few training instances per class while
preserving the knowledge of the previously learned ones. SoftNet jointly learns
the model weights and adaptive non-binary soft masks at a base training session
in which each mask consists of the major and minor subnetwork; the former aims
to minimize catastrophic forgetting during training, and the latter aims to
avoid overfitting to a few samples in each new training session. We provide
comprehensive empirical validations demonstrating that our SoftNet effectively
tackles the few-shot incremental learning problem by surpassing the performance
of state-of-the-art baselines over benchmark datasets.
|
2209.07817 | Sangseon Lee | Sangseon Lee, Dohoon Lee, Yinhua Piao, Sun Kim | SPGP: Structure Prototype Guided Graph Pooling | 20 pages, 6 figures | NeurIPS 2022 Workshop GLFrontiers | null | null | cs.LG cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While graph neural networks (GNNs) have been successful for node
classification tasks and link prediction tasks in graph, learning graph-level
representations still remains a challenge. For the graph-level representation,
it is important to learn both representation of neighboring nodes, i.e.,
aggregation, and graph structural information. A number of graph pooling
methods have been developed for this goal. However, most of the existing
pooling methods utilize k-hop neighborhood without considering explicit
structural information in a graph. In this paper, we propose Structure
Prototype Guided Pooling (SPGP) that utilizes prior graph structures to
overcome the limitation. SPGP formulates graph structures as learnable
prototype vectors and computes the affinity between nodes and prototype
vectors. This leads to a novel node scoring scheme that prioritizes informative
nodes while encapsulating the useful structures of the graph. Our experimental
results show that SPGP outperforms state-of-the-art graph pooling methods on
graph classification benchmark datasets in both accuracy and scalability.
| [
{
"version": "v1",
"created": "Fri, 16 Sep 2022 09:33:09 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 07:38:13 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Lee",
"Sangseon",
""
],
[
"Lee",
"Dohoon",
""
],
[
"Piao",
"Yinhua",
""
],
[
"Kim",
"Sun",
""
]
] | TITLE: SPGP: Structure Prototype Guided Graph Pooling
ABSTRACT: While graph neural networks (GNNs) have been successful for node
classification tasks and link prediction tasks in graph, learning graph-level
representations still remains a challenge. For the graph-level representation,
it is important to learn both representation of neighboring nodes, i.e.,
aggregation, and graph structural information. A number of graph pooling
methods have been developed for this goal. However, most of the existing
pooling methods utilize k-hop neighborhood without considering explicit
structural information in a graph. In this paper, we propose Structure
Prototype Guided Pooling (SPGP) that utilizes prior graph structures to
overcome the limitation. SPGP formulates graph structures as learnable
prototype vectors and computes the affinity between nodes and prototype
vectors. This leads to a novel node scoring scheme that prioritizes informative
nodes while encapsulating the useful structures of the graph. Our experimental
results show that SPGP outperforms state-of-the-art graph pooling methods on
graph classification benchmark datasets in both accuracy and scalability.
|
2209.08196 | Jordan Ford | Jeff Ford and Jordan Ford | Lossless SIMD Compression of LiDAR Range and Attribute Scan Sequences | null | null | null | null | cs.RO cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | As LiDAR sensors have become ubiquitous, the need for an efficient LiDAR data
compression algorithm has increased. Modern LiDARs produce gigabytes of scan
data per hour and are often used in applications with limited compute,
bandwidth, and storage resources.
We present a fast, lossless compression algorithm for LiDAR range and
attribute scan sequences including multiple-return range, signal, reflectivity,
and ambient infrared. Our algorithm -- dubbed "Jiffy" -- achieves substantial
compression by exploiting spatiotemporal redundancy and sparsity. Speed is
accomplished by maximizing use of single-instruction-multiple-data (SIMD)
instructions. In autonomous driving, infrastructure monitoring, drone
inspection, and handheld mapping benchmarks, the Jiffy algorithm consistently
outcompresses competing lossless codecs while operating at speeds in excess of
65M points/sec on a single core. In a typical autonomous vehicle use case,
single-threaded Jiffy achieves 6x compression of centimeter-precision range
scans at 500+ scans per second. To ensure reproducibility and enable adoption,
the software is freely available as an open source library.
| [
{
"version": "v1",
"created": "Fri, 16 Sep 2022 23:29:48 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 21:30:32 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Ford",
"Jeff",
""
],
[
"Ford",
"Jordan",
""
]
] | TITLE: Lossless SIMD Compression of LiDAR Range and Attribute Scan Sequences
ABSTRACT: As LiDAR sensors have become ubiquitous, the need for an efficient LiDAR data
compression algorithm has increased. Modern LiDARs produce gigabytes of scan
data per hour and are often used in applications with limited compute,
bandwidth, and storage resources.
We present a fast, lossless compression algorithm for LiDAR range and
attribute scan sequences including multiple-return range, signal, reflectivity,
and ambient infrared. Our algorithm -- dubbed "Jiffy" -- achieves substantial
compression by exploiting spatiotemporal redundancy and sparsity. Speed is
accomplished by maximizing use of single-instruction-multiple-data (SIMD)
instructions. In autonomous driving, infrastructure monitoring, drone
inspection, and handheld mapping benchmarks, the Jiffy algorithm consistently
outcompresses competing lossless codecs while operating at speeds in excess of
65M points/sec on a single core. In a typical autonomous vehicle use case,
single-threaded Jiffy achieves 6x compression of centimeter-precision range
scans at 500+ scans per second. To ensure reproducibility and enable adoption,
the software is freely available as an open source library.
|
2209.08343 | Mihnea-Alexandru Tomita | Mihnea-Alexandru Tomita, Bruno Ferrarini, Michael Milford, Klaus
McDonald-Maier, Shoaib Ehsan | Data Efficient Visual Place Recognition Using Extremely JPEG-Compressed
Images | The paper is currently under-review. 8 pages, 8 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual Place Recognition (VPR) is the ability of a robotic platform to
correctly interpret visual stimuli from its on-board cameras in order to
determine whether it is currently located in a previously visited place,
despite different viewpoint, illumination and appearance changes. JPEG is a
widely used image compression standard that is capable of significantly
reducing the size of an image at the cost of image clarity. For applications
where several robotic platforms are simultaneously deployed, the visual data
gathered must be transmitted remotely between each robot. Hence, JPEG
compression can be employed to drastically reduce the amount of data
transmitted over a communication channel, as working with limited bandwidth for
VPR can be proven to be a challenging task. However, the effects of JPEG
compression on the performance of current VPR techniques have not been
previously studied. For this reason, this paper presents an in-depth study of
JPEG compression in VPR related scenarios. We use a selection of
well-established VPR techniques on well-established benchmark datasets with
various amounts of compression applied. We show that by introducing
compression, the VPR performance is drastically reduced, especially in the
higher spectrum of compression. Moreover, this paper demonstrates how
fine-tuning a CNN can be utilised as an optimisation method for JPEG compressed
data to perform more consistently with the image transformations detected in
extremely JPEG compressed images.
| [
{
"version": "v1",
"created": "Sat, 17 Sep 2022 14:46:28 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 13:13:47 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Tomita",
"Mihnea-Alexandru",
""
],
[
"Ferrarini",
"Bruno",
""
],
[
"Milford",
"Michael",
""
],
[
"McDonald-Maier",
"Klaus",
""
],
[
"Ehsan",
"Shoaib",
""
]
] | TITLE: Data Efficient Visual Place Recognition Using Extremely JPEG-Compressed
Images
ABSTRACT: Visual Place Recognition (VPR) is the ability of a robotic platform to
correctly interpret visual stimuli from its on-board cameras in order to
determine whether it is currently located in a previously visited place,
despite different viewpoint, illumination and appearance changes. JPEG is a
widely used image compression standard that is capable of significantly
reducing the size of an image at the cost of image clarity. For applications
where several robotic platforms are simultaneously deployed, the visual data
gathered must be transmitted remotely between each robot. Hence, JPEG
compression can be employed to drastically reduce the amount of data
transmitted over a communication channel, as working with limited bandwidth for
VPR can be proven to be a challenging task. However, the effects of JPEG
compression on the performance of current VPR techniques have not been
previously studied. For this reason, this paper presents an in-depth study of
JPEG compression in VPR related scenarios. We use a selection of
well-established VPR techniques on well-established benchmark datasets with
various amounts of compression applied. We show that by introducing
compression, the VPR performance is drastically reduced, especially in the
higher spectrum of compression. Moreover, this paper demonstrates how
fine-tuning a CNN can be utilised as an optimisation method for JPEG compressed
data to perform more consistently with the image transformations detected in
extremely JPEG compressed images.
|
2209.13036 | Guangyao Zhai | Guangyao Zhai, Dianye Huang, Shun-Cheng Wu, Hyunjun Jung, Yan Di,
Fabian Manhardt, Federico Tombari, Nassir Navab and Benjamin Busam | MonoGraspNet: 6-DoF Grasping with a Single RGB Image | ICRA 2023 accepted. Project website:
https://sites.google.com/view/monograsp | null | null | null | cs.RO cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | 6-DoF robotic grasping is a long-lasting but unsolved problem. Recent methods
utilize strong 3D networks to extract geometric grasping representations from
depth sensors, demonstrating superior accuracy on common objects but perform
unsatisfactorily on photometrically challenging objects, e.g., objects in
transparent or reflective materials. The bottleneck lies in that the surface of
these objects can not reflect back accurate depth due to the absorption or
refraction of light. In this paper, in contrast to exploiting the inaccurate
depth data, we propose the first RGB-only 6-DoF grasping pipeline called
MonoGraspNet that utilizes stable 2D features to simultaneously handle
arbitrary object grasping and overcome the problems induced by photometrically
challenging objects. MonoGraspNet leverages keypoint heatmap and normal map to
recover the 6-DoF grasping poses represented by our novel representation
parameterized with 2D keypoints with corresponding depth, grasping direction,
grasping width, and angle. Extensive experiments in real scenes demonstrate
that our method can achieve competitive results in grasping common objects and
surpass the depth-based competitor by a large margin in grasping
photometrically challenging objects. To further stimulate robotic manipulation
research, we additionally annotate and open-source a multi-view and multi-scene
real-world grasping dataset, containing 120 objects of mixed photometric
complexity with 20M accurate grasping labels.
| [
{
"version": "v1",
"created": "Mon, 26 Sep 2022 21:29:50 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 15:27:29 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Zhai",
"Guangyao",
""
],
[
"Huang",
"Dianye",
""
],
[
"Wu",
"Shun-Cheng",
""
],
[
"Jung",
"Hyunjun",
""
],
[
"Di",
"Yan",
""
],
[
"Manhardt",
"Fabian",
""
],
[
"Tombari",
"Federico",
""
],
[
"Navab",
"Nassir",
""
],
[
"Busam",
"Benjamin",
""
]
] | TITLE: MonoGraspNet: 6-DoF Grasping with a Single RGB Image
ABSTRACT: 6-DoF robotic grasping is a long-lasting but unsolved problem. Recent methods
utilize strong 3D networks to extract geometric grasping representations from
depth sensors, demonstrating superior accuracy on common objects but perform
unsatisfactorily on photometrically challenging objects, e.g., objects in
transparent or reflective materials. The bottleneck lies in that the surface of
these objects can not reflect back accurate depth due to the absorption or
refraction of light. In this paper, in contrast to exploiting the inaccurate
depth data, we propose the first RGB-only 6-DoF grasping pipeline called
MonoGraspNet that utilizes stable 2D features to simultaneously handle
arbitrary object grasping and overcome the problems induced by photometrically
challenging objects. MonoGraspNet leverages keypoint heatmap and normal map to
recover the 6-DoF grasping poses represented by our novel representation
parameterized with 2D keypoints with corresponding depth, grasping direction,
grasping width, and angle. Extensive experiments in real scenes demonstrate
that our method can achieve competitive results in grasping common objects and
surpass the depth-based competitor by a large margin in grasping
photometrically challenging objects. To further stimulate robotic manipulation
research, we additionally annotate and open-source a multi-view and multi-scene
real-world grasping dataset, containing 120 objects of mixed photometric
complexity with 20M accurate grasping labels.
|
2209.13657 | Neelay Joglekar | Neelay Joglekar, Fei Liu, Ryan Orosco, Michael Yip | Suture Thread Spline Reconstruction from Endoscopic Images for Robotic
Surgery with Reliability-driven Keypoint Detection | To be published in ICRA 2023 | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automating the process of manipulating and delivering sutures during robotic
surgery is a prominent problem at the frontier of surgical robotics, as
automating this task can significantly reduce surgeons' fatigue during
tele-operated surgery and allow them to spend more time addressing higher-level
clinical decision making. Accomplishing autonomous suturing and suture
manipulation in the real world requires accurate suture thread localization and
reconstruction, the process of creating a 3D shape representation of suture
thread from 2D stereo camera surgical image pairs. This is a very challenging
problem due to how limited pixel information is available for the threads, as
well as their sensitivity to lighting and specular reflection. We present a
suture thread reconstruction work that uses reliable keypoints and a Minimum
Variation Spline (MVS) smoothing optimization to construct a 3D centerline from
a segmented surgical image pair. This method is comparable to previous suture
thread reconstruction works, with the possible benefit of increased accuracy of
grasping point estimation. Our code and datasets will be available at:
https://github.com/ucsdarclab/thread-reconstruction.
| [
{
"version": "v1",
"created": "Tue, 27 Sep 2022 19:48:20 GMT"
},
{
"version": "v2",
"created": "Sun, 16 Oct 2022 10:12:01 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Feb 2023 22:42:12 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Joglekar",
"Neelay",
""
],
[
"Liu",
"Fei",
""
],
[
"Orosco",
"Ryan",
""
],
[
"Yip",
"Michael",
""
]
] | TITLE: Suture Thread Spline Reconstruction from Endoscopic Images for Robotic
Surgery with Reliability-driven Keypoint Detection
ABSTRACT: Automating the process of manipulating and delivering sutures during robotic
surgery is a prominent problem at the frontier of surgical robotics, as
automating this task can significantly reduce surgeons' fatigue during
tele-operated surgery and allow them to spend more time addressing higher-level
clinical decision making. Accomplishing autonomous suturing and suture
manipulation in the real world requires accurate suture thread localization and
reconstruction, the process of creating a 3D shape representation of suture
thread from 2D stereo camera surgical image pairs. This is a very challenging
problem due to how limited pixel information is available for the threads, as
well as their sensitivity to lighting and specular reflection. We present a
suture thread reconstruction work that uses reliable keypoints and a Minimum
Variation Spline (MVS) smoothing optimization to construct a 3D centerline from
a segmented surgical image pair. This method is comparable to previous suture
thread reconstruction works, with the possible benefit of increased accuracy of
grasping point estimation. Our code and datasets will be available at:
https://github.com/ucsdarclab/thread-reconstruction.
|
2209.14935 | Yann Ollivier | Ahmed Touati, J\'er\'emy Rapin, Yann Ollivier | Does Zero-Shot Reinforcement Learning Exist? | null | International Conference on Learning Representations, 2023 | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A zero-shot RL agent is an agent that can solve any RL task in a given
environment, instantly with no additional planning or learning, after an
initial reward-free learning phase. This marks a shift from the reward-centric
RL paradigm towards "controllable" agents that can follow arbitrary
instructions in an environment. Current RL agents can solve families of related
tasks at best, or require planning anew for each task. Strategies for
approximate zero-shot RL ave been suggested using successor features (SFs)
[BBQ+ 18] or forward-backward (FB) representations [TO21], but testing has been
limited.
After clarifying the relationships between these schemes, we introduce
improved losses and new SF models, and test the viability of zero-shot RL
schemes systematically on tasks from the Unsupervised RL benchmark [LYL+21]. To
disentangle universal representation learning from exploration, we work in an
offline setting and repeat the tests on several existing replay buffers.
SFs appear to suffer from the choice of the elementary state features. SFs
with Laplacian eigenfunctions do well, while SFs based on auto-encoders,
inverse curiosity, transition models, low-rank transition matrix, contrastive
learning, or diversity (APS), perform unconsistently. In contrast, FB
representations jointly learn the elementary and successor features from a
single, principled criterion. They perform best and consistently across the
board, reaching 85% of supervised RL performance with a good replay buffer, in
a zero-shot manner.
| [
{
"version": "v1",
"created": "Thu, 29 Sep 2022 16:54:05 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 18:01:09 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Touati",
"Ahmed",
""
],
[
"Rapin",
"Jérémy",
""
],
[
"Ollivier",
"Yann",
""
]
] | TITLE: Does Zero-Shot Reinforcement Learning Exist?
ABSTRACT: A zero-shot RL agent is an agent that can solve any RL task in a given
environment, instantly with no additional planning or learning, after an
initial reward-free learning phase. This marks a shift from the reward-centric
RL paradigm towards "controllable" agents that can follow arbitrary
instructions in an environment. Current RL agents can solve families of related
tasks at best, or require planning anew for each task. Strategies for
approximate zero-shot RL ave been suggested using successor features (SFs)
[BBQ+ 18] or forward-backward (FB) representations [TO21], but testing has been
limited.
After clarifying the relationships between these schemes, we introduce
improved losses and new SF models, and test the viability of zero-shot RL
schemes systematically on tasks from the Unsupervised RL benchmark [LYL+21]. To
disentangle universal representation learning from exploration, we work in an
offline setting and repeat the tests on several existing replay buffers.
SFs appear to suffer from the choice of the elementary state features. SFs
with Laplacian eigenfunctions do well, while SFs based on auto-encoders,
inverse curiosity, transition models, low-rank transition matrix, contrastive
learning, or diversity (APS), perform unconsistently. In contrast, FB
representations jointly learn the elementary and successor features from a
single, principled criterion. They perform best and consistently across the
board, reaching 85% of supervised RL performance with a good replay buffer, in
a zero-shot manner.
|
2210.00120 | Ruiqi Ni | Ruiqi Ni, Ahmed H. Qureshi | NTFields: Neural Time Fields for Physics-Informed Robot Motion Planning | null | null | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural Motion Planners (NMPs) have emerged as a promising tool for solving
robot navigation tasks in complex environments. However, these methods often
require expert data for learning, which limits their application to scenarios
where data generation is time-consuming. Recent developments have also led to
physics-informed deep neural models capable of representing complex dynamical
Partial Differential Equations (PDEs). Inspired by these developments, we
propose Neural Time Fields (NTFields) for robot motion planning in cluttered
scenarios. Our framework represents a wave propagation model generating
continuous arrival time to find path solutions informed by a nonlinear
first-order PDE called Eikonal Equation. We evaluate our method in various
cluttered 3D environments, including the Gibson dataset, and demonstrate its
ability to solve motion planning problems for 4-DOF and 6-DOF robot
manipulators where the traditional grid-based Eikonal planners often face the
curse of dimensionality. Furthermore, the results show that our method exhibits
high success rates and significantly lower computational times than the
state-of-the-art methods, including NMPs that require training data from
classical planners.
| [
{
"version": "v1",
"created": "Fri, 30 Sep 2022 22:34:54 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 15:23:49 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Ni",
"Ruiqi",
""
],
[
"Qureshi",
"Ahmed H.",
""
]
] | TITLE: NTFields: Neural Time Fields for Physics-Informed Robot Motion Planning
ABSTRACT: Neural Motion Planners (NMPs) have emerged as a promising tool for solving
robot navigation tasks in complex environments. However, these methods often
require expert data for learning, which limits their application to scenarios
where data generation is time-consuming. Recent developments have also led to
physics-informed deep neural models capable of representing complex dynamical
Partial Differential Equations (PDEs). Inspired by these developments, we
propose Neural Time Fields (NTFields) for robot motion planning in cluttered
scenarios. Our framework represents a wave propagation model generating
continuous arrival time to find path solutions informed by a nonlinear
first-order PDE called Eikonal Equation. We evaluate our method in various
cluttered 3D environments, including the Gibson dataset, and demonstrate its
ability to solve motion planning problems for 4-DOF and 6-DOF robot
manipulators where the traditional grid-based Eikonal planners often face the
curse of dimensionality. Furthermore, the results show that our method exhibits
high success rates and significantly lower computational times than the
state-of-the-art methods, including NMPs that require training data from
classical planners.
|
2210.00312 | Ningyu Zhang | Ningyu Zhang, Lei Li, Xiang Chen, Xiaozhuan Liang, Shumin Deng, Huajun
Chen | Multimodal Analogical Reasoning over Knowledge Graphs | Accepted by ICLR 2023. The project website is
https://zjunlp.github.io/project/MKG_Analogy/introduction.html | null | null | null | cs.CL cs.AI cs.CV cs.LG cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Analogical reasoning is fundamental to human cognition and holds an important
place in various fields. However, previous studies mainly focus on single-modal
analogical reasoning and ignore taking advantage of structure knowledge.
Notably, the research in cognitive psychology has demonstrated that information
from multimodal sources always brings more powerful cognitive transfer than
single modality sources. To this end, we introduce the new task of multimodal
analogical reasoning over knowledge graphs, which requires multimodal reasoning
ability with the help of background knowledge. Specifically, we construct a
Multimodal Analogical Reasoning dataSet (MARS) and a multimodal knowledge graph
MarKG. We evaluate with multimodal knowledge graph embedding and pre-trained
Transformer baselines, illustrating the potential challenges of the proposed
task. We further propose a novel model-agnostic Multimodal analogical reasoning
framework with Transformer (MarT) motivated by the structure mapping theory,
which can obtain better performance. Code and datasets are available in
https://github.com/zjunlp/MKG_Analogy.
| [
{
"version": "v1",
"created": "Sat, 1 Oct 2022 16:24:15 GMT"
},
{
"version": "v2",
"created": "Tue, 29 Nov 2022 10:40:00 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jan 2023 05:26:39 GMT"
},
{
"version": "v4",
"created": "Wed, 1 Mar 2023 02:51:12 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Zhang",
"Ningyu",
""
],
[
"Li",
"Lei",
""
],
[
"Chen",
"Xiang",
""
],
[
"Liang",
"Xiaozhuan",
""
],
[
"Deng",
"Shumin",
""
],
[
"Chen",
"Huajun",
""
]
] | TITLE: Multimodal Analogical Reasoning over Knowledge Graphs
ABSTRACT: Analogical reasoning is fundamental to human cognition and holds an important
place in various fields. However, previous studies mainly focus on single-modal
analogical reasoning and ignore taking advantage of structure knowledge.
Notably, the research in cognitive psychology has demonstrated that information
from multimodal sources always brings more powerful cognitive transfer than
single modality sources. To this end, we introduce the new task of multimodal
analogical reasoning over knowledge graphs, which requires multimodal reasoning
ability with the help of background knowledge. Specifically, we construct a
Multimodal Analogical Reasoning dataSet (MARS) and a multimodal knowledge graph
MarKG. We evaluate with multimodal knowledge graph embedding and pre-trained
Transformer baselines, illustrating the potential challenges of the proposed
task. We further propose a novel model-agnostic Multimodal analogical reasoning
framework with Transformer (MarT) motivated by the structure mapping theory,
which can obtain better performance. Code and datasets are available in
https://github.com/zjunlp/MKG_Analogy.
|
2210.02875 | Tianbao Xie | Zhoujun Cheng, Tianbao Xie, Peng Shi, Chengzu Li, Rahul Nadkarni,
Yushi Hu, Caiming Xiong, Dragomir Radev, Mari Ostendorf, Luke Zettlemoyer,
Noah A. Smith, Tao Yu | Binding Language Models in Symbolic Languages | ICLR 2023 camera ready, 27 pages, 10 figures | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Though end-to-end neural approaches have recently been dominating NLP tasks
in both performance and ease-of-use, they lack interpretability and robustness.
We propose Binder, a training-free neural-symbolic framework that maps the task
input to a program, which (1) allows binding a unified API of language model
(LM) functionalities to a programming language (e.g., SQL, Python) to extend
its grammar coverage and thus tackle more diverse questions, (2) adopts an LM
as both the program parser and the underlying model called by the API during
execution, and (3) requires only a few in-context exemplar annotations.
Specifically, we employ GPT-3 Codex as the LM. In the parsing stage, with only
a few in-context exemplars, Codex is able to identify the part of the task
input that cannot be answerable by the original programming language, correctly
generate API calls to prompt Codex to solve the unanswerable part, and identify
where to place the API calls while being compatible with the original grammar.
In the execution stage, Codex can perform versatile functionalities (e.g.,
commonsense QA, information extraction) given proper prompts in the API calls.
Binder achieves state-of-the-art results on WikiTableQuestions and TabFact
datasets, with explicit output programs that benefit human debugging. Note that
previous best systems are all finetuned on tens of thousands of task-specific
samples, while Binder only uses dozens of annotations as in-context exemplars
without any training. Our code is available at https://github.com/HKUNLP/Binder .
| [
{
"version": "v1",
"created": "Thu, 6 Oct 2022 12:55:17 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 03:21:40 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Cheng",
"Zhoujun",
""
],
[
"Xie",
"Tianbao",
""
],
[
"Shi",
"Peng",
""
],
[
"Li",
"Chengzu",
""
],
[
"Nadkarni",
"Rahul",
""
],
[
"Hu",
"Yushi",
""
],
[
"Xiong",
"Caiming",
""
],
[
"Radev",
"Dragomir",
""
],
[
"Ostendorf",
"Mari",
""
],
[
"Zettlemoyer",
"Luke",
""
],
[
"Smith",
"Noah A.",
""
],
[
"Yu",
"Tao",
""
]
] | TITLE: Binding Language Models in Symbolic Languages
ABSTRACT: Though end-to-end neural approaches have recently been dominating NLP tasks
in both performance and ease-of-use, they lack interpretability and robustness.
We propose Binder, a training-free neural-symbolic framework that maps the task
input to a program, which (1) allows binding a unified API of language model
(LM) functionalities to a programming language (e.g., SQL, Python) to extend
its grammar coverage and thus tackle more diverse questions, (2) adopts an LM
as both the program parser and the underlying model called by the API during
execution, and (3) requires only a few in-context exemplar annotations.
Specifically, we employ GPT-3 Codex as the LM. In the parsing stage, with only
a few in-context exemplars, Codex is able to identify the part of the task
input that cannot be answerable by the original programming language, correctly
generate API calls to prompt Codex to solve the unanswerable part, and identify
where to place the API calls while being compatible with the original grammar.
In the execution stage, Codex can perform versatile functionalities (e.g.,
commonsense QA, information extraction) given proper prompts in the API calls.
Binder achieves state-of-the-art results on WikiTableQuestions and TabFact
datasets, with explicit output programs that benefit human debugging. Note that
previous best systems are all finetuned on tens of thousands of task-specific
samples, while Binder only uses dozens of annotations as in-context exemplars
without any training. Our code is available at https://github.com/HKUNLP/Binder .
|
2210.07494 | Keyu Duan | Keyu Duan, Zirui Liu, Peihao Wang, Wenqing Zheng, Kaixiong Zhou,
Tianlong Chen, Xia Hu, Zhangyang Wang | A Comprehensive Study on Large-Scale Graph Training: Benchmarking and
Rethinking | Accepted by NeurIPS 2022 Dataset and Benchmark Track | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large-scale graph training is a notoriously challenging problem for graph
neural networks (GNNs). Due to the nature of evolving graph structures into the
training process, vanilla GNNs usually fail to scale up, limited by the GPU
memory space. Up to now, though numerous scalable GNN architectures have been
proposed, we still lack a comprehensive survey and fair benchmark of this
reservoir to find the rationale for designing scalable GNNs. To this end, we
first systematically formulate the representative methods of large-scale graph
training into several branches and further establish a fair and consistent
benchmark for them by a greedy hyperparameter searching. In addition, regarding
efficiency, we theoretically evaluate the time and space complexity of various
branches and empirically compare them w.r.t GPU memory usage, throughput, and
convergence. Furthermore, We analyze the pros and cons for various branches of
scalable GNNs and then present a new ensembling training manner, named EnGCN,
to address the existing issues. Our code is available at
https://github.com/VITA-Group/Large_Scale_GCN_Benchmarking.
| [
{
"version": "v1",
"created": "Fri, 14 Oct 2022 03:43:05 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 08:02:27 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Duan",
"Keyu",
""
],
[
"Liu",
"Zirui",
""
],
[
"Wang",
"Peihao",
""
],
[
"Zheng",
"Wenqing",
""
],
[
"Zhou",
"Kaixiong",
""
],
[
"Chen",
"Tianlong",
""
],
[
"Hu",
"Xia",
""
],
[
"Wang",
"Zhangyang",
""
]
] | TITLE: A Comprehensive Study on Large-Scale Graph Training: Benchmarking and
Rethinking
ABSTRACT: Large-scale graph training is a notoriously challenging problem for graph
neural networks (GNNs). Due to the nature of evolving graph structures into the
training process, vanilla GNNs usually fail to scale up, limited by the GPU
memory space. Up to now, though numerous scalable GNN architectures have been
proposed, we still lack a comprehensive survey and fair benchmark of this
reservoir to find the rationale for designing scalable GNNs. To this end, we
first systematically formulate the representative methods of large-scale graph
training into several branches and further establish a fair and consistent
benchmark for them by a greedy hyperparameter searching. In addition, regarding
efficiency, we theoretically evaluate the time and space complexity of various
branches and empirically compare them w.r.t GPU memory usage, throughput, and
convergence. Furthermore, We analyze the pros and cons for various branches of
scalable GNNs and then present a new ensembling training manner, named EnGCN,
to address the existing issues. Our code is available at
https://github.com/VITA-Group/Large_Scale_GCN_Benchmarking.
|
2210.11636 | Mona Wang | Mona Wang and Jonathan Mayer | Self-Censorship Under Law: A Case Study of the Hong Kong National
Security Law | null | Privacy Enhancing Technologies, Free and Open Communications on
the Internet. Issue 1. (2023) 46-55 | null | null | cs.CY | http://creativecommons.org/licenses/by/4.0/ | We study how legislation that restricts speech can induce online
self-censorship and alter online discourse, using the recent Hong Kong national
security law as a case study. We collect a dataset of 7 million historical
Tweets from Hong Kong users, supplemented with historical snapshots of Tweet
streams collected by other researchers. We compare online activity before and
after enactment of the national security law, and we find that Hong Kong users
demonstrate two types of self-censorship. First, Hong Kong users are more
likely than a control group, sampled randomly from historical snapshots of
Tweet streams, to remove past online activity. Specifically, Hong Kong users
are over a third more likely than the control group to delete or restrict their
account and over twice as likely to delete past posts. Second, we find that
Hong Kong users post less often about politically sensitive topics that have
been censored on social media in mainland China. This trend continues to
increase.
| [
{
"version": "v1",
"created": "Thu, 20 Oct 2022 23:42:04 GMT"
},
{
"version": "v2",
"created": "Wed, 15 Feb 2023 15:12:14 GMT"
},
{
"version": "v3",
"created": "Fri, 17 Feb 2023 16:29:56 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Wang",
"Mona",
""
],
[
"Mayer",
"Jonathan",
""
]
] | TITLE: Self-Censorship Under Law: A Case Study of the Hong Kong National
Security Law
ABSTRACT: We study how legislation that restricts speech can induce online
self-censorship and alter online discourse, using the recent Hong Kong national
security law as a case study. We collect a dataset of 7 million historical
Tweets from Hong Kong users, supplemented with historical snapshots of Tweet
streams collected by other researchers. We compare online activity before and
after enactment of the national security law, and we find that Hong Kong users
demonstrate two types of self-censorship. First, Hong Kong users are more
likely than a control group, sampled randomly from historical snapshots of
Tweet streams, to remove past online activity. Specifically, Hong Kong users
are over a third more likely than the control group to delete or restrict their
account and over twice as likely to delete past posts. Second, we find that
Hong Kong users post less often about politically sensitive topics that have
been censored on social media in mainland China. This trend continues to
increase.
|
2210.15274 | Yudong Chen | Yudong Chen, Sen Wang, Jiajun Liu, Xuwei Xu, Frank de Hoog, Zi Huang | Improved Feature Distillation via Projector Ensemble | NeurIPS 2022 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In knowledge distillation, previous feature distillation methods mainly focus
on the design of loss functions and the selection of the distilled layers,
while the effect of the feature projector between the student and the teacher
remains under-explored. In this paper, we first discuss a plausible mechanism
of the projector with empirical evidence and then propose a new feature
distillation method based on a projector ensemble for further performance
improvement. We observe that the student network benefits from a projector even
if the feature dimensions of the student and the teacher are the same. Training
a student backbone without a projector can be considered as a multi-task
learning process, namely achieving discriminative feature extraction for
classification and feature matching between the student and the teacher for
distillation at the same time. We hypothesize and empirically verify that
without a projector, the student network tends to overfit the teacher's feature
distributions despite having different architecture and weights initialization.
This leads to degradation on the quality of the student's deep features that
are eventually used in classification. Adding a projector, on the other hand,
disentangles the two learning tasks and helps the student network to focus
better on the main feature extraction task while still being able to utilize
teacher features as a guidance through the projector. Motivated by the positive
effect of the projector in feature distillation, we propose an ensemble of
projectors to further improve the quality of student features. Experimental
results on different datasets with a series of teacher-student pairs illustrate
the effectiveness of the proposed method.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2022 09:08:40 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 02:06:06 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Chen",
"Yudong",
""
],
[
"Wang",
"Sen",
""
],
[
"Liu",
"Jiajun",
""
],
[
"Xu",
"Xuwei",
""
],
[
"de Hoog",
"Frank",
""
],
[
"Huang",
"Zi",
""
]
] | TITLE: Improved Feature Distillation via Projector Ensemble
ABSTRACT: In knowledge distillation, previous feature distillation methods mainly focus
on the design of loss functions and the selection of the distilled layers,
while the effect of the feature projector between the student and the teacher
remains under-explored. In this paper, we first discuss a plausible mechanism
of the projector with empirical evidence and then propose a new feature
distillation method based on a projector ensemble for further performance
improvement. We observe that the student network benefits from a projector even
if the feature dimensions of the student and the teacher are the same. Training
a student backbone without a projector can be considered as a multi-task
learning process, namely achieving discriminative feature extraction for
classification and feature matching between the student and the teacher for
distillation at the same time. We hypothesize and empirically verify that
without a projector, the student network tends to overfit the teacher's feature
distributions despite having different architecture and weights initialization.
This leads to degradation on the quality of the student's deep features that
are eventually used in classification. Adding a projector, on the other hand,
disentangles the two learning tasks and helps the student network to focus
better on the main feature extraction task while still being able to utilize
teacher features as a guidance through the projector. Motivated by the positive
effect of the projector in feature distillation, we propose an ensemble of
projectors to further improve the quality of student features. Experimental
results on different datasets with a series of teacher-student pairs illustrate
the effectiveness of the proposed method.
|
2210.17028 | Thy Nguyen | Thy Nguyen, Anamay Chaturvedi, Huy L\^e Nguyen | Improved Learning-augmented Algorithms for k-means and k-medians
Clustering | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the problem of clustering in the learning-augmented setting,
where we are given a data set in $d$-dimensional Euclidean space, and a label
for each data point given by an oracle indicating what subsets of points should
be clustered together. This setting captures situations where we have access to
some auxiliary information about the data set relevant for our clustering
objective, for instance the labels output by a neural network. Following prior
work, we assume that there are at most an $\alpha \in (0,c)$ for some $c<1$
fraction of false positives and false negatives in each predicted cluster, in
the absence of which the labels would attain the optimal clustering cost
$\mathrm{OPT}$.
For a dataset of size $m$, we propose a deterministic $k$-means algorithm
that produces centers with improved bound on clustering cost compared to the
previous randomized algorithm while preserving the $O( d m \log m)$ runtime.
Furthermore, our algorithm works even when the predictions are not very
accurate, i.e. our bound holds for $\alpha$ up to $1/2$, an improvement over
$\alpha$ being at most $1/7$ in the previous work. For the $k$-medians problem
we improve upon prior work by achieving a biquadratic improvement in the
dependence of the approximation factor on the accuracy parameter $\alpha$ to
get a cost of $(1+O(\alpha))\mathrm{OPT}$, while requiring essentially just
$O(md \log^3 m/\alpha)$ runtime.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2022 03:00:11 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 17:41:25 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Nguyen",
"Thy",
""
],
[
"Chaturvedi",
"Anamay",
""
],
[
"Nguyen",
"Huy Lê",
""
]
] | TITLE: Improved Learning-augmented Algorithms for k-means and k-medians
Clustering
ABSTRACT: We consider the problem of clustering in the learning-augmented setting,
where we are given a data set in $d$-dimensional Euclidean space, and a label
for each data point given by an oracle indicating what subsets of points should
be clustered together. This setting captures situations where we have access to
some auxiliary information about the data set relevant for our clustering
objective, for instance the labels output by a neural network. Following prior
work, we assume that there are at most an $\alpha \in (0,c)$ for some $c<1$
fraction of false positives and false negatives in each predicted cluster, in
the absence of which the labels would attain the optimal clustering cost
$\mathrm{OPT}$.
For a dataset of size $m$, we propose a deterministic $k$-means algorithm
that produces centers with improved bound on clustering cost compared to the
previous randomized algorithm while preserving the $O( d m \log m)$ runtime.
Furthermore, our algorithm works even when the predictions are not very
accurate, i.e. our bound holds for $\alpha$ up to $1/2$, an improvement over
$\alpha$ being at most $1/7$ in the previous work. For the $k$-medians problem
we improve upon prior work by achieving a biquadratic improvement in the
dependence of the approximation factor on the accuracy parameter $\alpha$ to
get a cost of $(1+O(\alpha))\mathrm{OPT}$, while requiring essentially just
$O(md \log^3 m/\alpha)$ runtime.
|
2210.17390 | John Musgrave | John Musgrave, Temesguen Messay-Kebede, David Kapp, Anca Ralescu | Latent Semantic Structure in Malicious Programs | 11 pages, 5 figures | Springer Nature Switzerland, Communications in Computer and
Information Science (CCIS, volume 1761), MDIS 2022, pp 234,246 | 10.1007/978-3-031-27034-5_16 | null | cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Latent Semantic Analysis is a method of matrix decomposition used for
discovering topics and topic weights in natural language documents. This study
uses Latent Semantic Analysis to analyze the composition of binaries of
malicious programs. The semantic representation of the term frequency vector
representation yields a set of topics, each topic being a composition of terms.
The vectors and topics were evaluated quantitatively using a spatial
representation. This semantic analysis provides a more abstract representation
of the program derived from its term frequency analysis. We use a metric space
to represent a program as a collection of vectors, and a distance metric to
evaluate their similarity within a topic. The segmentation of the vectors in
this dataset provides increased resolution into the program structure.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2022 15:16:28 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Musgrave",
"John",
""
],
[
"Messay-Kebede",
"Temesguen",
""
],
[
"Kapp",
"David",
""
],
[
"Ralescu",
"Anca",
""
]
] | TITLE: Latent Semantic Structure in Malicious Programs
ABSTRACT: Latent Semantic Analysis is a method of matrix decomposition used for
discovering topics and topic weights in natural language documents. This study
uses Latent Semantic Analysis to analyze the composition of binaries of
malicious programs. The semantic representation of the term frequency vector
representation yields a set of topics, each topic being a composition of terms.
The vectors and topics were evaluated quantitatively using a spatial
representation. This semantic analysis provides a more abstract representation
of the program derived from its term frequency analysis. We use a metric space
to represent a program as a collection of vectors, and a distance metric to
evaluate their similarity within a topic. The segmentation of the vectors in
this dataset provides increased resolution into the program structure.
|
2211.04717 | Yu Chen | Yu Chen, Wen Ding, Junjie Lai | Improving Noisy Student Training on Non-target Domain Data for Automatic
Speech Recognition | This paper is accepted by the ICASSP 2023 conference | null | null | null | cs.SD cs.CL eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Noisy Student Training (NST) has recently demonstrated extremely strong
performance in Automatic Speech Recognition(ASR). In this paper, we propose a
data selection strategy named LM Filter to improve the performance of NST on
non-target domain data in ASR tasks. Hypotheses with and without a Language
Model are generated and the CER differences between them are utilized as a
filter threshold. Results reveal that significant improvements of 10.4%
compared with no data filtering baselines. We can achieve 3.31% CER in
AISHELL-1 test set, which is best result from our knowledge without any other
supervised data. We also perform evaluations on the supervised 1000 hour
AISHELL-2 dataset and competitive results of 4.73% CER can be achieved.
| [
{
"version": "v1",
"created": "Wed, 9 Nov 2022 07:23:15 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 17:59:06 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Chen",
"Yu",
""
],
[
"Ding",
"Wen",
""
],
[
"Lai",
"Junjie",
""
]
] | TITLE: Improving Noisy Student Training on Non-target Domain Data for Automatic
Speech Recognition
ABSTRACT: Noisy Student Training (NST) has recently demonstrated extremely strong
performance in Automatic Speech Recognition(ASR). In this paper, we propose a
data selection strategy named LM Filter to improve the performance of NST on
non-target domain data in ASR tasks. Hypotheses with and without a Language
Model are generated and the CER differences between them are utilized as a
filter threshold. Results reveal that significant improvements of 10.4%
compared with no data filtering baselines. We can achieve 3.31% CER in
AISHELL-1 test set, which is best result from our knowledge without any other
supervised data. We also perform evaluations on the supervised 1000 hour
AISHELL-2 dataset and competitive results of 4.73% CER can be achieved.
|
2211.06545 | Jianan Zhao | Jianan Zhao, Qianlong Wen, Mingxuan Ju, Chuxu Zhang, Yanfang Ye | Self-Supervised Graph Structure Refinement for Graph Neural Networks | WSDM 2023 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Graph structure learning (GSL), which aims to learn the adjacency matrix for
graph neural networks (GNNs), has shown great potential in boosting the
performance of GNNs. Most existing GSL works apply a joint learning framework
where the estimated adjacency matrix and GNN parameters are optimized for
downstream tasks. However, as GSL is essentially a link prediction task, whose
goal may largely differ from the goal of the downstream task. The inconsistency
of these two goals limits the GSL methods to learn the potential optimal graph
structure. Moreover, the joint learning framework suffers from scalability
issues in terms of time and space during the process of estimation and
optimization of the adjacency matrix. To mitigate these issues, we propose a
graph structure refinement (GSR) framework with a pretrain-finetune pipeline.
Specifically, The pre-training phase aims to comprehensively estimate the
underlying graph structure by a multi-view contrastive learning framework with
both intra- and inter-view link prediction tasks. Then, the graph structure is
refined by adding and removing edges according to the edge probabilities
estimated by the pre-trained model. Finally, the fine-tuning GNN is initialized
by the pre-trained model and optimized toward downstream tasks. With the
refined graph structure remaining static in the fine-tuning space, GSR avoids
estimating and optimizing graph structure in the fine-tuning phase which enjoys
great scalability and efficiency. Moreover, the fine-tuning GNN is boosted by
both migrating knowledge and refining graphs. Extensive experiments are
conducted to evaluate the effectiveness (best performance on six benchmark
datasets), efficiency, and scalability (13.8x faster using 32.8% GPU memory
compared to the best GSL baseline on Cora) of the proposed model.
| [
{
"version": "v1",
"created": "Sat, 12 Nov 2022 02:01:46 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Nov 2022 03:25:24 GMT"
},
{
"version": "v3",
"created": "Wed, 14 Dec 2022 16:01:03 GMT"
},
{
"version": "v4",
"created": "Wed, 1 Mar 2023 02:27:46 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Zhao",
"Jianan",
""
],
[
"Wen",
"Qianlong",
""
],
[
"Ju",
"Mingxuan",
""
],
[
"Zhang",
"Chuxu",
""
],
[
"Ye",
"Yanfang",
""
]
] | TITLE: Self-Supervised Graph Structure Refinement for Graph Neural Networks
ABSTRACT: Graph structure learning (GSL), which aims to learn the adjacency matrix for
graph neural networks (GNNs), has shown great potential in boosting the
performance of GNNs. Most existing GSL works apply a joint learning framework
where the estimated adjacency matrix and GNN parameters are optimized for
downstream tasks. However, as GSL is essentially a link prediction task, whose
goal may largely differ from the goal of the downstream task. The inconsistency
of these two goals limits the GSL methods to learn the potential optimal graph
structure. Moreover, the joint learning framework suffers from scalability
issues in terms of time and space during the process of estimation and
optimization of the adjacency matrix. To mitigate these issues, we propose a
graph structure refinement (GSR) framework with a pretrain-finetune pipeline.
Specifically, The pre-training phase aims to comprehensively estimate the
underlying graph structure by a multi-view contrastive learning framework with
both intra- and inter-view link prediction tasks. Then, the graph structure is
refined by adding and removing edges according to the edge probabilities
estimated by the pre-trained model. Finally, the fine-tuning GNN is initialized
by the pre-trained model and optimized toward downstream tasks. With the
refined graph structure remaining static in the fine-tuning space, GSR avoids
estimating and optimizing graph structure in the fine-tuning phase which enjoys
great scalability and efficiency. Moreover, the fine-tuning GNN is boosted by
both migrating knowledge and refining graphs. Extensive experiments are
conducted to evaluate the effectiveness (best performance on six benchmark
datasets), efficiency, and scalability (13.8x faster using 32.8% GPU memory
compared to the best GSL baseline on Cora) of the proposed model.
|
2211.07711 | Meimei Wu | Junyi He, Meimei Wu, Meng Li, Xiaobo Zhu, Feng Ye | Multilevel Transformer For Multimodal Emotion Recognition | 5 pages, 2 figures, 3 tables. Submitted to ICASSP 2023 | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal emotion recognition has attracted much attention recently. Fusing
multiple modalities effectively with limited labeled data is a challenging
task. Considering the success of pre-trained model and fine-grained nature of
emotion expression, it is reasonable to take these two aspects into
consideration. Unlike previous methods that mainly focus on one aspect, we
introduce a novel multi-granularity framework, which combines fine-grained
representation with pre-trained utterance-level representation. Inspired by
Transformer TTS, we propose a multilevel transformer model to perform
fine-grained multimodal emotion recognition. Specifically, we explore different
methods to incorporate phoneme-level embedding with word-level embedding. To
perform multi-granularity learning, we simply combine multilevel transformer
model with Albert. Extensive experimental results show that both our multilevel
transformer model and multi-granularity model outperform previous
state-of-the-art approaches on IEMOCAP dataset with text transcripts and speech
signal.
| [
{
"version": "v1",
"created": "Wed, 26 Oct 2022 10:31:24 GMT"
},
{
"version": "v2",
"created": "Wed, 16 Nov 2022 06:56:41 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"He",
"Junyi",
""
],
[
"Wu",
"Meimei",
""
],
[
"Li",
"Meng",
""
],
[
"Zhu",
"Xiaobo",
""
],
[
"Ye",
"Feng",
""
]
] | TITLE: Multilevel Transformer For Multimodal Emotion Recognition
ABSTRACT: Multimodal emotion recognition has attracted much attention recently. Fusing
multiple modalities effectively with limited labeled data is a challenging
task. Considering the success of pre-trained model and fine-grained nature of
emotion expression, it is reasonable to take these two aspects into
consideration. Unlike previous methods that mainly focus on one aspect, we
introduce a novel multi-granularity framework, which combines fine-grained
representation with pre-trained utterance-level representation. Inspired by
Transformer TTS, we propose a multilevel transformer model to perform
fine-grained multimodal emotion recognition. Specifically, we explore different
methods to incorporate phoneme-level embedding with word-level embedding. To
perform multi-granularity learning, we simply combine multilevel transformer
model with Albert. Extensive experimental results show that both our multilevel
transformer model and multi-granularity model outperform previous
state-of-the-art approaches on IEMOCAP dataset with text transcripts and speech
signal.
|
2211.11853 | Adri\'an Javaloy | Adri\'an Javaloy, Pablo Sanchez-Martin, Amit Levi and Isabel Valera | Learnable Graph Convolutional Attention Networks | Published as a conference paper at ICLR 2023. 35 pages, 5 figures | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing Graph Neural Networks (GNNs) compute the message exchange between
nodes by either aggregating uniformly (convolving) the features of all the
neighboring nodes, or by applying a non-uniform score (attending) to the
features. Recent works have shown the strengths and weaknesses of the resulting
GNN architectures, respectively, GCNs and GATs. In this work, we aim at
exploiting the strengths of both approaches to their full extent. To this end,
we first introduce the graph convolutional attention layer (CAT), which relies
on convolutions to compute the attention scores. Unfortunately, as in the case
of GCNs and GATs, we show that there exists no clear winner between the three
(neither theoretically nor in practice) as their performance directly depends
on the nature of the data (i.e., of the graph and features). This result brings
us to the main contribution of our work, the learnable graph convolutional
attention network (L-CAT): a GNN architecture that automatically interpolates
between GCN, GAT and CAT in each layer, by adding only two scalar parameters.
Our results demonstrate that L-CAT is able to efficiently combine different GNN
layers along the network, outperforming competing methods in a wide range of
datasets, and resulting in a more robust model that reduces the need of
cross-validating.
| [
{
"version": "v1",
"created": "Mon, 21 Nov 2022 21:08:58 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 20:33:42 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Javaloy",
"Adrián",
""
],
[
"Sanchez-Martin",
"Pablo",
""
],
[
"Levi",
"Amit",
""
],
[
"Valera",
"Isabel",
""
]
] | TITLE: Learnable Graph Convolutional Attention Networks
ABSTRACT: Existing Graph Neural Networks (GNNs) compute the message exchange between
nodes by either aggregating uniformly (convolving) the features of all the
neighboring nodes, or by applying a non-uniform score (attending) to the
features. Recent works have shown the strengths and weaknesses of the resulting
GNN architectures, respectively, GCNs and GATs. In this work, we aim at
exploiting the strengths of both approaches to their full extent. To this end,
we first introduce the graph convolutional attention layer (CAT), which relies
on convolutions to compute the attention scores. Unfortunately, as in the case
of GCNs and GATs, we show that there exists no clear winner between the three
(neither theoretically nor in practice) as their performance directly depends
on the nature of the data (i.e., of the graph and features). This result brings
us to the main contribution of our work, the learnable graph convolutional
attention network (L-CAT): a GNN architecture that automatically interpolates
between GCN, GAT and CAT in each layer, by adding only two scalar parameters.
Our results demonstrate that L-CAT is able to efficiently combine different GNN
layers along the network, outperforming competing methods in a wide range of
datasets, and resulting in a more robust model that reduces the need of
cross-validating.
|
2211.12500 | Ankan Kumar Bhunia | Ankan Kumar Bhunia, Salman Khan, Hisham Cholakkal, Rao Muhammad Anwer,
Jorma Laaksonen, Mubarak Shah, Fahad Shahbaz Khan | Person Image Synthesis via Denoising Diffusion Model | Accepted to CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The pose-guided person image generation task requires synthesizing
photorealistic images of humans in arbitrary poses. The existing approaches use
generative adversarial networks that do not necessarily maintain realistic
textures or need dense correspondences that struggle to handle complex
deformations and severe occlusions. In this work, we show how denoising
diffusion models can be applied for high-fidelity person image synthesis with
strong sample diversity and enhanced mode coverage of the learnt data
distribution. Our proposed Person Image Diffusion Model (PIDM) disintegrates
the complex transfer problem into a series of simpler forward-backward
denoising steps. This helps in learning plausible source-to-target
transformation trajectories that result in faithful textures and undistorted
appearance details. We introduce a 'texture diffusion module' based on
cross-attention to accurately model the correspondences between appearance and
pose information available in source and target images. Further, we propose
'disentangled classifier-free guidance' to ensure close resemblance between the
conditional inputs and the synthesized output in terms of both pose and
appearance information. Our extensive results on two large-scale benchmarks and
a user study demonstrate the photorealism of our proposed approach under
challenging scenarios. We also show how our generated images can help in
downstream tasks. Our code and models will be publicly released.
| [
{
"version": "v1",
"created": "Tue, 22 Nov 2022 18:59:50 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 14:17:39 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Bhunia",
"Ankan Kumar",
""
],
[
"Khan",
"Salman",
""
],
[
"Cholakkal",
"Hisham",
""
],
[
"Anwer",
"Rao Muhammad",
""
],
[
"Laaksonen",
"Jorma",
""
],
[
"Shah",
"Mubarak",
""
],
[
"Khan",
"Fahad Shahbaz",
""
]
] | TITLE: Person Image Synthesis via Denoising Diffusion Model
ABSTRACT: The pose-guided person image generation task requires synthesizing
photorealistic images of humans in arbitrary poses. The existing approaches use
generative adversarial networks that do not necessarily maintain realistic
textures or need dense correspondences that struggle to handle complex
deformations and severe occlusions. In this work, we show how denoising
diffusion models can be applied for high-fidelity person image synthesis with
strong sample diversity and enhanced mode coverage of the learnt data
distribution. Our proposed Person Image Diffusion Model (PIDM) disintegrates
the complex transfer problem into a series of simpler forward-backward
denoising steps. This helps in learning plausible source-to-target
transformation trajectories that result in faithful textures and undistorted
appearance details. We introduce a 'texture diffusion module' based on
cross-attention to accurately model the correspondences between appearance and
pose information available in source and target images. Further, we propose
'disentangled classifier-free guidance' to ensure close resemblance between the
conditional inputs and the synthesized output in terms of both pose and
appearance information. Our extensive results on two large-scale benchmarks and
a user study demonstrate the photorealism of our proposed approach under
challenging scenarios. We also show how our generated images can help in
downstream tasks. Our code and models will be publicly released.
|
2211.15612 | Qi Tian | Qi Tian, Kun Kuang, Furui Liu, Baoxiang Wang | Learning from Good Trajectories in Offline Multi-Agent Reinforcement
Learning | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Offline multi-agent reinforcement learning (MARL) aims to learn effective
multi-agent policies from pre-collected datasets, which is an important step
toward the deployment of multi-agent systems in real-world applications.
However, in practice, each individual behavior policy that generates
multi-agent joint trajectories usually has a different level of how well it
performs. e.g., an agent is a random policy while other agents are medium
policies. In the cooperative game with global reward, one agent learned by
existing offline MARL often inherits this random policy, jeopardizing the
performance of the entire team. In this paper, we investigate offline MARL with
explicit consideration on the diversity of agent-wise trajectories and propose
a novel framework called Shared Individual Trajectories (SIT) to address this
problem. Specifically, an attention-based reward decomposition network assigns
the credit to each agent through a differentiable key-value memory mechanism in
an offline manner. These decomposed credits are then used to reconstruct the
joint offline datasets into prioritized experience replay with individual
trajectories, thereafter agents can share their good trajectories and
conservatively train their policies with a graph attention network (GAT) based
critic. We evaluate our method in both discrete control (i.e., StarCraft II and
multi-agent particle environment) and continuous control (i.e, multi-agent
mujoco). The results indicate that our method achieves significantly better
results in complex and mixed offline multi-agent datasets, especially when the
difference of data quality between individual trajectories is large.
| [
{
"version": "v1",
"created": "Mon, 28 Nov 2022 18:11:26 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 14:48:03 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Tian",
"Qi",
""
],
[
"Kuang",
"Kun",
""
],
[
"Liu",
"Furui",
""
],
[
"Wang",
"Baoxiang",
""
]
] | TITLE: Learning from Good Trajectories in Offline Multi-Agent Reinforcement
Learning
ABSTRACT: Offline multi-agent reinforcement learning (MARL) aims to learn effective
multi-agent policies from pre-collected datasets, which is an important step
toward the deployment of multi-agent systems in real-world applications.
However, in practice, each individual behavior policy that generates
multi-agent joint trajectories usually has a different level of how well it
performs. e.g., an agent is a random policy while other agents are medium
policies. In the cooperative game with global reward, one agent learned by
existing offline MARL often inherits this random policy, jeopardizing the
performance of the entire team. In this paper, we investigate offline MARL with
explicit consideration on the diversity of agent-wise trajectories and propose
a novel framework called Shared Individual Trajectories (SIT) to address this
problem. Specifically, an attention-based reward decomposition network assigns
the credit to each agent through a differentiable key-value memory mechanism in
an offline manner. These decomposed credits are then used to reconstruct the
joint offline datasets into prioritized experience replay with individual
trajectories, thereafter agents can share their good trajectories and
conservatively train their policies with a graph attention network (GAT) based
critic. We evaluate our method in both discrete control (i.e., StarCraft II and
multi-agent particle environment) and continuous control (i.e, multi-agent
mujoco). The results indicate that our method achieves significantly better
results in complex and mixed offline multi-agent datasets, especially when the
difference of data quality between individual trajectories is large.
|
2212.00959 | Jinhao Jiang | Jinhao Jiang, Kun Zhou, Wayne Xin Zhao and Ji-Rong Wen | UniKGQA: Unified Retrieval and Reasoning for Solving Multi-hop Question
Answering Over Knowledge Graph | Camera-ready of ICLR 2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-hop Question Answering over Knowledge Graph~(KGQA) aims to find the
answer entities that are multiple hops away from the topic entities mentioned
in a natural language question on a large-scale Knowledge Graph (KG). To cope
with the vast search space, existing work usually adopts a two-stage approach:
it first retrieves a relatively small subgraph related to the question and then
performs the reasoning on the subgraph to find the answer entities accurately.
Although these two stages are highly related, previous work employs very
different technical solutions for developing the retrieval and reasoning
models, neglecting their relatedness in task essence. In this paper, we propose
UniKGQA, a novel approach for multi-hop KGQA task, by unifying retrieval and
reasoning in both model architecture and parameter learning. For model
architecture, UniKGQA consists of a semantic matching module based on a
pre-trained language model~(PLM) for question-relation semantic matching, and a
matching information propagation module to propagate the matching information
along the directed edges on KGs. For parameter learning, we design a shared
pre-training task based on question-relation matching for both retrieval and
reasoning models, and then propose retrieval- and reasoning-oriented
fine-tuning strategies. Compared with previous studies, our approach is more
unified, tightly relating the retrieval and reasoning stages. Extensive
experiments on three benchmark datasets have demonstrated the effectiveness of
our method on the multi-hop KGQA task. Our codes and data are publicly
available at~\url{https://github.com/RUCAIBox/UniKGQA}.
| [
{
"version": "v1",
"created": "Fri, 2 Dec 2022 04:08:09 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 07:48:46 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Jiang",
"Jinhao",
""
],
[
"Zhou",
"Kun",
""
],
[
"Zhao",
"Wayne Xin",
""
],
[
"Wen",
"Ji-Rong",
""
]
] | TITLE: UniKGQA: Unified Retrieval and Reasoning for Solving Multi-hop Question
Answering Over Knowledge Graph
ABSTRACT: Multi-hop Question Answering over Knowledge Graph~(KGQA) aims to find the
answer entities that are multiple hops away from the topic entities mentioned
in a natural language question on a large-scale Knowledge Graph (KG). To cope
with the vast search space, existing work usually adopts a two-stage approach:
it first retrieves a relatively small subgraph related to the question and then
performs the reasoning on the subgraph to find the answer entities accurately.
Although these two stages are highly related, previous work employs very
different technical solutions for developing the retrieval and reasoning
models, neglecting their relatedness in task essence. In this paper, we propose
UniKGQA, a novel approach for multi-hop KGQA task, by unifying retrieval and
reasoning in both model architecture and parameter learning. For model
architecture, UniKGQA consists of a semantic matching module based on a
pre-trained language model~(PLM) for question-relation semantic matching, and a
matching information propagation module to propagate the matching information
along the directed edges on KGs. For parameter learning, we design a shared
pre-training task based on question-relation matching for both retrieval and
reasoning models, and then propose retrieval- and reasoning-oriented
fine-tuning strategies. Compared with previous studies, our approach is more
unified, tightly relating the retrieval and reasoning stages. Extensive
experiments on three benchmark datasets have demonstrated the effectiveness of
our method on the multi-hop KGQA task. Our codes and data are publicly
available at~\url{https://github.com/RUCAIBox/UniKGQA}.
|
2212.01433 | Sheng Liu | Sheng Liu, Xu Zhang, Nitesh Sekhar, Yue Wu, Prateek Singhal, Carlos
Fernandez-Granda | Avoiding spurious correlations via logit correction | 17 pages, 6 figures | null | null | null | cs.LG cs.CL cs.CV eess.IV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Empirical studies suggest that machine learning models trained with empirical
risk minimization (ERM) often rely on attributes that may be spuriously
correlated with the class labels. Such models typically lead to poor
performance during inference for data lacking such correlations. In this work,
we explicitly consider a situation where potential spurious correlations are
present in the majority of training data. In contrast with existing approaches,
which use the ERM model outputs to detect the samples without spurious
correlations and either heuristically upweight or upsample those samples, we
propose the logit correction (LC) loss, a simple yet effective improvement on
the softmax cross-entropy loss, to correct the sample logit. We demonstrate
that minimizing the LC loss is equivalent to maximizing the group-balanced
accuracy, so the proposed LC could mitigate the negative impacts of spurious
correlations. Our extensive experimental results further reveal that the
proposed LC loss outperforms state-of-the-art solutions on multiple popular
benchmarks by a large margin, an average 5.5\% absolute improvement, without
access to spurious attribute labels. LC is also competitive with oracle methods
that make use of the attribute labels. Code is available at
https://github.com/shengliu66/LC.
| [
{
"version": "v1",
"created": "Fri, 2 Dec 2022 20:30:59 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 23:06:09 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Liu",
"Sheng",
""
],
[
"Zhang",
"Xu",
""
],
[
"Sekhar",
"Nitesh",
""
],
[
"Wu",
"Yue",
""
],
[
"Singhal",
"Prateek",
""
],
[
"Fernandez-Granda",
"Carlos",
""
]
] | TITLE: Avoiding spurious correlations via logit correction
ABSTRACT: Empirical studies suggest that machine learning models trained with empirical
risk minimization (ERM) often rely on attributes that may be spuriously
correlated with the class labels. Such models typically lead to poor
performance during inference for data lacking such correlations. In this work,
we explicitly consider a situation where potential spurious correlations are
present in the majority of training data. In contrast with existing approaches,
which use the ERM model outputs to detect the samples without spurious
correlations and either heuristically upweight or upsample those samples, we
propose the logit correction (LC) loss, a simple yet effective improvement on
the softmax cross-entropy loss, to correct the sample logit. We demonstrate
that minimizing the LC loss is equivalent to maximizing the group-balanced
accuracy, so the proposed LC could mitigate the negative impacts of spurious
correlations. Our extensive experimental results further reveal that the
proposed LC loss outperforms state-of-the-art solutions on multiple popular
benchmarks by a large margin, an average 5.5\% absolute improvement, without
access to spurious attribute labels. LC is also competitive with oracle methods
that make use of the attribute labels. Code is available at
https://github.com/shengliu66/LC.
|
2212.02742 | Tom Ginsberg | Tom Ginsberg, Zhongyuan Liang, and Rahul G. Krishnan | A Learning Based Hypothesis Test for Harmful Covariate Shift | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The ability to quickly and accurately identify covariate shift at test time
is a critical and often overlooked component of safe machine learning systems
deployed in high-risk domains. While methods exist for detecting when
predictions should not be made on out-of-distribution test examples,
identifying distributional level differences between training and test time can
help determine when a model should be removed from the deployment setting and
retrained. In this work, we define harmful covariate shift (HCS) as a change in
distribution that may weaken the generalization of a predictive model. To
detect HCS, we use the discordance between an ensemble of classifiers trained
to agree on training data and disagree on test data. We derive a loss function
for training this ensemble and show that the disagreement rate and entropy
represent powerful discriminative statistics for HCS. Empirically, we
demonstrate the ability of our method to detect harmful covariate shift with
statistical certainty on a variety of high-dimensional datasets. Across
numerous domains and modalities, we show state-of-the-art performance compared
to existing methods, particularly when the number of observed test samples is
small.
| [
{
"version": "v1",
"created": "Tue, 6 Dec 2022 04:15:24 GMT"
},
{
"version": "v2",
"created": "Wed, 7 Dec 2022 03:19:24 GMT"
},
{
"version": "v3",
"created": "Tue, 24 Jan 2023 17:15:33 GMT"
},
{
"version": "v4",
"created": "Tue, 31 Jan 2023 18:38:44 GMT"
},
{
"version": "v5",
"created": "Fri, 17 Feb 2023 21:47:56 GMT"
},
{
"version": "v6",
"created": "Wed, 1 Mar 2023 18:19:45 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Ginsberg",
"Tom",
""
],
[
"Liang",
"Zhongyuan",
""
],
[
"Krishnan",
"Rahul G.",
""
]
] | TITLE: A Learning Based Hypothesis Test for Harmful Covariate Shift
ABSTRACT: The ability to quickly and accurately identify covariate shift at test time
is a critical and often overlooked component of safe machine learning systems
deployed in high-risk domains. While methods exist for detecting when
predictions should not be made on out-of-distribution test examples,
identifying distributional level differences between training and test time can
help determine when a model should be removed from the deployment setting and
retrained. In this work, we define harmful covariate shift (HCS) as a change in
distribution that may weaken the generalization of a predictive model. To
detect HCS, we use the discordance between an ensemble of classifiers trained
to agree on training data and disagree on test data. We derive a loss function
for training this ensemble and show that the disagreement rate and entropy
represent powerful discriminative statistics for HCS. Empirically, we
demonstrate the ability of our method to detect harmful covariate shift with
statistical certainty on a variety of high-dimensional datasets. Across
numerous domains and modalities, we show state-of-the-art performance compared
to existing methods, particularly when the number of observed test samples is
small.
|
2212.10898 | Khai Loong Aw | Khai Loong Aw, Mariya Toneva | Training language models to summarize narratives improves brain
alignment | ICLR 2023 (notable top 25%) | null | null | null | cs.CL q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Building systems that achieve a deeper understanding of language is one of
the central goals of natural language processing (NLP). Towards this goal,
recent works have begun to train language models on narrative datasets which
require extracting the most critical information by integrating across long
contexts. However, it is still an open question whether these models are
learning a deeper understanding of the text, or if the models are simply
learning a heuristic to complete the task. This work investigates this further
by turning to the one language processing system that truly understands complex
language: the human brain. We show that training language models for deeper
narrative understanding results in richer representations that have improved
alignment to human brain activity. We further find that the improvements in
brain alignment are larger for character names than for other discourse
features, which indicates that these models are learning important narrative
elements. Taken together, these results suggest that this type of training can
indeed lead to deeper language understanding. These findings have consequences
both for cognitive neuroscience by revealing some of the significant factors
behind brain-NLP alignment, and for NLP by highlighting that understanding of
long-range context can be improved beyond language modeling.
| [
{
"version": "v1",
"created": "Wed, 21 Dec 2022 10:15:19 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 03:20:08 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Aw",
"Khai Loong",
""
],
[
"Toneva",
"Mariya",
""
]
] | TITLE: Training language models to summarize narratives improves brain
alignment
ABSTRACT: Building systems that achieve a deeper understanding of language is one of
the central goals of natural language processing (NLP). Towards this goal,
recent works have begun to train language models on narrative datasets which
require extracting the most critical information by integrating across long
contexts. However, it is still an open question whether these models are
learning a deeper understanding of the text, or if the models are simply
learning a heuristic to complete the task. This work investigates this further
by turning to the one language processing system that truly understands complex
language: the human brain. We show that training language models for deeper
narrative understanding results in richer representations that have improved
alignment to human brain activity. We further find that the improvements in
brain alignment are larger for character names than for other discourse
features, which indicates that these models are learning important narrative
elements. Taken together, these results suggest that this type of training can
indeed lead to deeper language understanding. These findings have consequences
both for cognitive neuroscience by revealing some of the significant factors
behind brain-NLP alignment, and for NLP by highlighting that understanding of
long-range context can be improved beyond language modeling.
|
2301.02328 | Joey Hejna | Divyansh Garg, Joey Hejna, Matthieu Geist, Stefano Ermon | Extreme Q-Learning: MaxEnt RL without Entropy | ICLR 2023 Oral | null | null | null | cs.LG cs.AI cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern Deep Reinforcement Learning (RL) algorithms require estimates of the
maximal Q-value, which are difficult to compute in continuous domains with an
infinite number of possible actions. In this work, we introduce a new update
rule for online and offline RL which directly models the maximal value using
Extreme Value Theory (EVT), drawing inspiration from economics. By doing so, we
avoid computing Q-values using out-of-distribution actions which is often a
substantial source of error. Our key insight is to introduce an objective that
directly estimates the optimal soft-value functions (LogSumExp) in the maximum
entropy RL setting without needing to sample from a policy. Using EVT, we
derive our \emph{Extreme Q-Learning} framework and consequently online and, for
the first time, offline MaxEnt Q-learning algorithms, that do not explicitly
require access to a policy or its entropy. Our method obtains consistently
strong performance in the D4RL benchmark, outperforming prior works by
\emph{10+ points} on the challenging Franka Kitchen tasks while offering
moderate improvements over SAC and TD3 on online DM Control tasks.
Visualizations and code can be found on our website at
https://div99.github.io/XQL/.
| [
{
"version": "v1",
"created": "Thu, 5 Jan 2023 23:14:38 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 22:14:18 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Garg",
"Divyansh",
""
],
[
"Hejna",
"Joey",
""
],
[
"Geist",
"Matthieu",
""
],
[
"Ermon",
"Stefano",
""
]
] | TITLE: Extreme Q-Learning: MaxEnt RL without Entropy
ABSTRACT: Modern Deep Reinforcement Learning (RL) algorithms require estimates of the
maximal Q-value, which are difficult to compute in continuous domains with an
infinite number of possible actions. In this work, we introduce a new update
rule for online and offline RL which directly models the maximal value using
Extreme Value Theory (EVT), drawing inspiration from economics. By doing so, we
avoid computing Q-values using out-of-distribution actions which is often a
substantial source of error. Our key insight is to introduce an objective that
directly estimates the optimal soft-value functions (LogSumExp) in the maximum
entropy RL setting without needing to sample from a policy. Using EVT, we
derive our \emph{Extreme Q-Learning} framework and consequently online and, for
the first time, offline MaxEnt Q-learning algorithms, that do not explicitly
require access to a policy or its entropy. Our method obtains consistently
strong performance in the D4RL benchmark, outperforming prior works by
\emph{10+ points} on the challenging Franka Kitchen tasks while offering
moderate improvements over SAC and TD3 on online DM Control tasks.
Visualizations and code can be found on our website at
https://div99.github.io/XQL/.
|
2301.12711 | Elmurod Kuriyozov | Maksud Sharipov, Elmurod Kuriyozov, Ollabergan Yuldashev, Ogabek
Sobirov | UzbekTagger: The rule-based POS tagger for Uzbek language | Preprint of the accepted paper to The 10th Language & Technology
Conference: Human Language Technologies as a Challenge for Computer Science
and Linguistics, April 21-23, 2023, Poznan, Poland | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This research paper presents a part-of-speech (POS) annotated dataset and
tagger tool for the low-resource Uzbek language. The dataset includes 12 tags,
which were used to develop a rule-based POS-tagger tool. The corpus text used
in the annotation process was made sure to be balanced over 20 different fields
in order to ensure its representativeness. Uzbek being an agglutinative
language so the most of the words in an Uzbek sentence are formed by adding
suffixes. This nature of it makes the POS-tagging task difficult to find the
stems of words and the right part-of-speech they belong to. The methodology
proposed in this research is the stemming of the words with an affix/suffix
stripping approach including database of the stem forms of the words in the
Uzbek language. The tagger tool was tested on the annotated dataset and showed
high accuracy in identifying and tagging parts of speech in Uzbek text. This
newly presented dataset and tagger tool can be used for a variety of natural
language processing tasks such as language modeling, machine translation, and
text-to-speech synthesis. The presented dataset is the first of its kind to be
made publicly available for Uzbek, and the POS-tagger tool created can also be
used as a pivot to use as a base for other closely-related Turkic languages.
| [
{
"version": "v1",
"created": "Mon, 30 Jan 2023 07:40:45 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 14:31:12 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Sharipov",
"Maksud",
""
],
[
"Kuriyozov",
"Elmurod",
""
],
[
"Yuldashev",
"Ollabergan",
""
],
[
"Sobirov",
"Ogabek",
""
]
] | TITLE: UzbekTagger: The rule-based POS tagger for Uzbek language
ABSTRACT: This research paper presents a part-of-speech (POS) annotated dataset and
tagger tool for the low-resource Uzbek language. The dataset includes 12 tags,
which were used to develop a rule-based POS-tagger tool. The corpus text used
in the annotation process was made sure to be balanced over 20 different fields
in order to ensure its representativeness. Uzbek being an agglutinative
language so the most of the words in an Uzbek sentence are formed by adding
suffixes. This nature of it makes the POS-tagging task difficult to find the
stems of words and the right part-of-speech they belong to. The methodology
proposed in this research is the stemming of the words with an affix/suffix
stripping approach including database of the stem forms of the words in the
Uzbek language. The tagger tool was tested on the annotated dataset and showed
high accuracy in identifying and tagging parts of speech in Uzbek text. This
newly presented dataset and tagger tool can be used for a variety of natural
language processing tasks such as language modeling, machine translation, and
text-to-speech synthesis. The presented dataset is the first of its kind to be
made publicly available for Uzbek, and the POS-tagger tool created can also be
used as a pivot to use as a base for other closely-related Turkic languages.
|
2301.13473 | Anima Majumder | Darshita Jain, Anima Majumder, Samrat Dutta and Swagat Kumar | CRC-RL: A Novel Visual Feature Representation Architecture for
Unsupervised Reinforcement Learning | null | null | null | null | cs.CV cs.AI cs.RO | http://creativecommons.org/licenses/by/4.0/ | This paper addresses the problem of visual feature representation learning
with an aim to improve the performance of end-to-end reinforcement learning
(RL) models. Specifically, a novel architecture is proposed that uses a
heterogeneous loss function, called CRC loss, to learn improved visual features
which can then be used for policy learning in RL. The CRC-loss function is a
combination of three individual loss functions, namely, contrastive,
reconstruction and consistency loss. The feature representation is learned in
parallel to the policy learning while sharing the weight updates through a
Siamese Twin encoder model. This encoder model is augmented with a decoder
network and a feature projection network to facilitate computation of the above
loss components. Through empirical analysis involving latent feature
visualization, an attempt is made to provide an insight into the role played by
this loss function in learning new action-dependent features and how they are
linked to the complexity of the problems being solved. The proposed
architecture, called CRC-RL, is shown to outperform the existing
state-of-the-art methods on the challenging Deep mind control suite
environments by a significant margin thereby creating a new benchmark in this
field.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2023 08:41:18 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 04:41:51 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Jain",
"Darshita",
""
],
[
"Majumder",
"Anima",
""
],
[
"Dutta",
"Samrat",
""
],
[
"Kumar",
"Swagat",
""
]
] | TITLE: CRC-RL: A Novel Visual Feature Representation Architecture for
Unsupervised Reinforcement Learning
ABSTRACT: This paper addresses the problem of visual feature representation learning
with an aim to improve the performance of end-to-end reinforcement learning
(RL) models. Specifically, a novel architecture is proposed that uses a
heterogeneous loss function, called CRC loss, to learn improved visual features
which can then be used for policy learning in RL. The CRC-loss function is a
combination of three individual loss functions, namely, contrastive,
reconstruction and consistency loss. The feature representation is learned in
parallel to the policy learning while sharing the weight updates through a
Siamese Twin encoder model. This encoder model is augmented with a decoder
network and a feature projection network to facilitate computation of the above
loss components. Through empirical analysis involving latent feature
visualization, an attempt is made to provide an insight into the role played by
this loss function in learning new action-dependent features and how they are
linked to the complexity of the problems being solved. The proposed
architecture, called CRC-RL, is shown to outperform the existing
state-of-the-art methods on the challenging Deep mind control suite
environments by a significant margin thereby creating a new benchmark in this
field.
|
2302.08639 | Mufan Sang | Mufan Sang, Yong Zhao, Gang Liu, John H.L. Hansen, Jian Wu | Improving Transformer-based Networks With Locality For Automatic Speaker
Verification | Accepted to ICASSP 2023 | null | null | null | eess.AS cs.LG cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, Transformer-based architectures have been explored for speaker
embedding extraction. Although the Transformer employs the self-attention
mechanism to efficiently model the global interaction between token embeddings,
it is inadequate for capturing short-range local context, which is essential
for the accurate extraction of speaker information. In this study, we enhance
the Transformer with the enhanced locality modeling in two directions. First,
we propose the Locality-Enhanced Conformer (LE-Confomer) by introducing
depth-wise convolution and channel-wise attention into the Conformer blocks.
Second, we present the Speaker Swin Transformer (SST) by adapting the Swin
Transformer, originally proposed for vision tasks, into speaker embedding
network. We evaluate the proposed approaches on the VoxCeleb datasets and a
large-scale Microsoft internal multilingual (MS-internal) dataset. The proposed
models achieve 0.75% EER on VoxCeleb 1 test set, outperforming the previously
proposed Transformer-based models and CNN-based models, such as ResNet34 and
ECAPA-TDNN. When trained on the MS-internal dataset, the proposed models
achieve promising results with 14.6% relative reduction in EER over the
Res2Net50 model.
| [
{
"version": "v1",
"created": "Fri, 17 Feb 2023 01:04:51 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 23:32:08 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Sang",
"Mufan",
""
],
[
"Zhao",
"Yong",
""
],
[
"Liu",
"Gang",
""
],
[
"Hansen",
"John H. L.",
""
],
[
"Wu",
"Jian",
""
]
] | TITLE: Improving Transformer-based Networks With Locality For Automatic Speaker
Verification
ABSTRACT: Recently, Transformer-based architectures have been explored for speaker
embedding extraction. Although the Transformer employs the self-attention
mechanism to efficiently model the global interaction between token embeddings,
it is inadequate for capturing short-range local context, which is essential
for the accurate extraction of speaker information. In this study, we enhance
the Transformer with the enhanced locality modeling in two directions. First,
we propose the Locality-Enhanced Conformer (LE-Confomer) by introducing
depth-wise convolution and channel-wise attention into the Conformer blocks.
Second, we present the Speaker Swin Transformer (SST) by adapting the Swin
Transformer, originally proposed for vision tasks, into speaker embedding
network. We evaluate the proposed approaches on the VoxCeleb datasets and a
large-scale Microsoft internal multilingual (MS-internal) dataset. The proposed
models achieve 0.75% EER on VoxCeleb 1 test set, outperforming the previously
proposed Transformer-based models and CNN-based models, such as ResNet34 and
ECAPA-TDNN. When trained on the MS-internal dataset, the proposed models
achieve promising results with 14.6% relative reduction in EER over the
Res2Net50 model.
|
2302.09369 | Bowen Lei | Bowen Lei, Ruqi Zhang, Dongkuan Xu, Bani Mallick | Calibrating the Rigged Lottery: Making All Tickets Reliable | null | null | null | null | cs.LG cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Although sparse training has been successfully used in various
resource-limited deep learning tasks to save memory, accelerate training, and
reduce inference time, the reliability of the produced sparse models remains
unexplored. Previous research has shown that deep neural networks tend to be
over-confident, and we find that sparse training exacerbates this problem.
Therefore, calibrating the sparse models is crucial for reliable prediction and
decision-making. In this paper, we propose a new sparse training method to
produce sparse models with improved confidence calibration. In contrast to
previous research that uses only one mask to control the sparse topology, our
method utilizes two masks, including a deterministic mask and a random mask.
The former efficiently searches and activates important weights by exploiting
the magnitude of weights and gradients. While the latter brings better
exploration and finds more appropriate weight values by random updates.
Theoretically, we prove our method can be viewed as a hierarchical variational
approximation of a probabilistic deep Gaussian process. Extensive experiments
on multiple datasets, model architectures, and sparsities show that our method
reduces ECE values by up to 47.8\% and simultaneously maintains or even
improves accuracy with only a slight increase in computation and storage
burden.
| [
{
"version": "v1",
"created": "Sat, 18 Feb 2023 15:53:55 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 03:48:17 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Lei",
"Bowen",
""
],
[
"Zhang",
"Ruqi",
""
],
[
"Xu",
"Dongkuan",
""
],
[
"Mallick",
"Bani",
""
]
] | TITLE: Calibrating the Rigged Lottery: Making All Tickets Reliable
ABSTRACT: Although sparse training has been successfully used in various
resource-limited deep learning tasks to save memory, accelerate training, and
reduce inference time, the reliability of the produced sparse models remains
unexplored. Previous research has shown that deep neural networks tend to be
over-confident, and we find that sparse training exacerbates this problem.
Therefore, calibrating the sparse models is crucial for reliable prediction and
decision-making. In this paper, we propose a new sparse training method to
produce sparse models with improved confidence calibration. In contrast to
previous research that uses only one mask to control the sparse topology, our
method utilizes two masks, including a deterministic mask and a random mask.
The former efficiently searches and activates important weights by exploiting
the magnitude of weights and gradients. While the latter brings better
exploration and finds more appropriate weight values by random updates.
Theoretically, we prove our method can be viewed as a hierarchical variational
approximation of a probabilistic deep Gaussian process. Extensive experiments
on multiple datasets, model architectures, and sparsities show that our method
reduces ECE values by up to 47.8\% and simultaneously maintains or even
improves accuracy with only a slight increase in computation and storage
burden.
|
2302.12360 | Igor L. Markov | Chung-Wei Lee, Pavlos Athanasios Apostolopulos, Igor L. Markov | Practical Knowledge Distillation: Using DNNs to Beat DNNs | 11 pages, 1 figure, 17 tables | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | For tabular data sets, we explore data and model distillation, as well as
data denoising. These techniques improve both gradient-boosting models and a
specialized DNN architecture. While gradient boosting is known to outperform
DNNs on tabular data, we close the gap for datasets with 100K+ rows and give
DNNs an advantage on small data sets. We extend these results with input-data
distillation and optimized ensembling to help DNN performance match or exceed
that of gradient boosting. As a theoretical justification of our practical
method, we prove its equivalence to classical cross-entropy knowledge
distillation. We also qualitatively explain the superiority of DNN ensembles
over XGBoost on small data sets. For an industry end-to-end real-time ML
platform with 4M production inferences per second, we develop a model-training
workflow based on data sampling that distills ensembles of models into a single
gradient-boosting model favored for high-performance real-time inference,
without performance loss. Empirical evaluation shows that the proposed
combination of methods consistently improves model accuracy over prior best
models across several production applications deployed worldwide.
| [
{
"version": "v1",
"created": "Thu, 23 Feb 2023 22:53:02 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 18:28:36 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Lee",
"Chung-Wei",
""
],
[
"Apostolopulos",
"Pavlos Athanasios",
""
],
[
"Markov",
"Igor L.",
""
]
] | TITLE: Practical Knowledge Distillation: Using DNNs to Beat DNNs
ABSTRACT: For tabular data sets, we explore data and model distillation, as well as
data denoising. These techniques improve both gradient-boosting models and a
specialized DNN architecture. While gradient boosting is known to outperform
DNNs on tabular data, we close the gap for datasets with 100K+ rows and give
DNNs an advantage on small data sets. We extend these results with input-data
distillation and optimized ensembling to help DNN performance match or exceed
that of gradient boosting. As a theoretical justification of our practical
method, we prove its equivalence to classical cross-entropy knowledge
distillation. We also qualitatively explain the superiority of DNN ensembles
over XGBoost on small data sets. For an industry end-to-end real-time ML
platform with 4M production inferences per second, we develop a model-training
workflow based on data sampling that distills ensembles of models into a single
gradient-boosting model favored for high-performance real-time inference,
without performance loss. Empirical evaluation shows that the proposed
combination of methods consistently improves model accuracy over prior best
models across several production applications deployed worldwide.
|
2302.12840 | Isabel Segura-Bedmar | Isabel Segura-Bedmar | HULAT at SemEval-2023 Task 10: Data augmentation for pre-trained
transformers applied to the detection of sexism in social media | The experiments are not reproducible because I did not use a seed for
replicability | null | null | null | cs.CL cs.AI cs.LG cs.NE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper describes our participation in SemEval-2023 Task 10, whose goal is
the detection of sexism in social media. We explore some of the most popular
transformer models such as BERT, DistilBERT, RoBERTa, and XLNet. We also study
different data augmentation techniques to increase the training dataset. During
the development phase, our best results were obtained by using RoBERTa and data
augmentation for tasks B and C. However, the use of synthetic data does not
improve the results for task C. We participated in the three subtasks. Our
approach still has much room for improvement, especially in the two
fine-grained classifications. All our code is available in the repository
https://github.com/isegura/hulat_edos.
| [
{
"version": "v1",
"created": "Fri, 24 Feb 2023 18:17:38 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 08:43:13 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Segura-Bedmar",
"Isabel",
""
]
] | TITLE: HULAT at SemEval-2023 Task 10: Data augmentation for pre-trained
transformers applied to the detection of sexism in social media
ABSTRACT: This paper describes our participation in SemEval-2023 Task 10, whose goal is
the detection of sexism in social media. We explore some of the most popular
transformer models such as BERT, DistilBERT, RoBERTa, and XLNet. We also study
different data augmentation techniques to increase the training dataset. During
the development phase, our best results were obtained by using RoBERTa and data
augmentation for tasks B and C. However, the use of synthetic data does not
improve the results for task C. We participated in the three subtasks. Our
approach still has much room for improvement, especially in the two
fine-grained classifications. All our code is available in the repository
https://github.com/isegura/hulat_edos.
|
2302.13277 | Siyuan Shen | Siyuan Shen, Feng Liu, Aimin Zhou | Mingling or Misalignment? Temporal Shift for Speech Emotion Recognition
with Pre-trained Representations | Accepted by 48th IEEE International Conference on Acoustics, Speech,
and Signal Processing (ICASSP 2023) | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fueled by recent advances of self-supervised models, pre-trained speech
representations proved effective for the downstream speech emotion recognition
(SER) task. Most prior works mainly focus on exploiting pre-trained
representations and just adopt a linear head on top of the pre-trained model,
neglecting the design of the downstream network. In this paper, we propose a
temporal shift module to mingle channel-wise information without introducing
any parameter or FLOP. With the temporal shift module, three designed baseline
building blocks evolve into corresponding shift variants, i.e. ShiftCNN,
ShiftLSTM, and Shiftformer. Moreover, to balance the trade-off between mingling
and misalignment, we propose two technical strategies, placement of shift and
proportion of shift. The family of temporal shift models all outperforms the
state-of-the-art methods on the benchmark IEMOCAP dataset under both finetuning
and feature extraction settings. Our code is available at
https://github.com/ECNU-Cross-Innovation-Lab/ShiftSER.
| [
{
"version": "v1",
"created": "Sun, 26 Feb 2023 09:35:24 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 06:02:14 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Shen",
"Siyuan",
""
],
[
"Liu",
"Feng",
""
],
[
"Zhou",
"Aimin",
""
]
] | TITLE: Mingling or Misalignment? Temporal Shift for Speech Emotion Recognition
with Pre-trained Representations
ABSTRACT: Fueled by recent advances of self-supervised models, pre-trained speech
representations proved effective for the downstream speech emotion recognition
(SER) task. Most prior works mainly focus on exploiting pre-trained
representations and just adopt a linear head on top of the pre-trained model,
neglecting the design of the downstream network. In this paper, we propose a
temporal shift module to mingle channel-wise information without introducing
any parameter or FLOP. With the temporal shift module, three designed baseline
building blocks evolve into corresponding shift variants, i.e. ShiftCNN,
ShiftLSTM, and Shiftformer. Moreover, to balance the trade-off between mingling
and misalignment, we propose two technical strategies, placement of shift and
proportion of shift. The family of temporal shift models all outperforms the
state-of-the-art methods on the benchmark IEMOCAP dataset under both finetuning
and feature extraction settings. Our code is available at
https://github.com/ECNU-Cross-Innovation-Lab/ShiftSER.
|
2302.13602 | Rahim Entezari | Rahim Entezari, Mitchell Wortsman, Olga Saukh, M.Moein Shariatnia,
Hanie Sedghi, Ludwig Schmidt | The Role of Pre-training Data in Transfer Learning | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The transfer learning paradigm of model pre-training and subsequent
fine-tuning produces high-accuracy models. While most studies recommend scaling
the pre-training size to benefit most from transfer learning, a question
remains: what data and method should be used for pre-training? We investigate
the impact of pre-training data distribution on the few-shot and full
fine-tuning performance using 3 pre-training methods (supervised, contrastive
language-image and image-image), 7 pre-training datasets, and 9 downstream
datasets. Through extensive controlled experiments, we find that the choice of
the pre-training data source is essential for the few-shot transfer, but its
role decreases as more data is made available for fine-tuning. Additionally, we
explore the role of data curation and examine the trade-offs between label
noise and the size of the pre-training dataset. We find that using 2000X more
pre-training data from LAION can match the performance of supervised ImageNet
pre-training. Furthermore, we investigate the effect of pre-training methods,
comparing language-image contrastive vs. image-image contrastive, and find that
the latter leads to better downstream accuracy
| [
{
"version": "v1",
"created": "Mon, 27 Feb 2023 09:10:08 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 13:48:55 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Entezari",
"Rahim",
""
],
[
"Wortsman",
"Mitchell",
""
],
[
"Saukh",
"Olga",
""
],
[
"Shariatnia",
"M. Moein",
""
],
[
"Sedghi",
"Hanie",
""
],
[
"Schmidt",
"Ludwig",
""
]
] | TITLE: The Role of Pre-training Data in Transfer Learning
ABSTRACT: The transfer learning paradigm of model pre-training and subsequent
fine-tuning produces high-accuracy models. While most studies recommend scaling
the pre-training size to benefit most from transfer learning, a question
remains: what data and method should be used for pre-training? We investigate
the impact of pre-training data distribution on the few-shot and full
fine-tuning performance using 3 pre-training methods (supervised, contrastive
language-image and image-image), 7 pre-training datasets, and 9 downstream
datasets. Through extensive controlled experiments, we find that the choice of
the pre-training data source is essential for the few-shot transfer, but its
role decreases as more data is made available for fine-tuning. Additionally, we
explore the role of data curation and examine the trade-offs between label
noise and the size of the pre-training dataset. We find that using 2000X more
pre-training data from LAION can match the performance of supervised ImageNet
pre-training. Furthermore, we investigate the effect of pre-training methods,
comparing language-image contrastive vs. image-image contrastive, and find that
the latter leads to better downstream accuracy
|
2302.13838 | Naoya Takahashi | Naoya Takahashi, Mayank K. Singh, Yuki Mitsufuji | Cross-modal Face- and Voice-style Transfer | null | null | null | null | cs.CV cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image-to-image translation and voice conversion enable the generation of a
new facial image and voice while maintaining some of the semantics such as a
pose in an image and linguistic content in audio, respectively. They can aid in
the content-creation process in many applications. However, as they are limited
to the conversion within each modality, matching the impression of the
generated face and voice remains an open question. We propose a cross-modal
style transfer framework called XFaVoT that jointly learns four tasks: image
translation and voice conversion tasks with audio or image guidance, which
enables the generation of ``face that matches given voice" and ``voice that
matches given face", and intra-modality translation tasks with a single
framework. Experimental results on multiple datasets show that XFaVoT achieves
cross-modal style translation of image and voice, outperforming baselines in
terms of quality, diversity, and face-voice correspondence.
| [
{
"version": "v1",
"created": "Mon, 27 Feb 2023 14:39:50 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 14:50:41 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Takahashi",
"Naoya",
""
],
[
"Singh",
"Mayank K.",
""
],
[
"Mitsufuji",
"Yuki",
""
]
] | TITLE: Cross-modal Face- and Voice-style Transfer
ABSTRACT: Image-to-image translation and voice conversion enable the generation of a
new facial image and voice while maintaining some of the semantics such as a
pose in an image and linguistic content in audio, respectively. They can aid in
the content-creation process in many applications. However, as they are limited
to the conversion within each modality, matching the impression of the
generated face and voice remains an open question. We propose a cross-modal
style transfer framework called XFaVoT that jointly learns four tasks: image
translation and voice conversion tasks with audio or image guidance, which
enables the generation of ``face that matches given voice" and ``voice that
matches given face", and intra-modality translation tasks with a single
framework. Experimental results on multiple datasets show that XFaVoT achieves
cross-modal style translation of image and voice, outperforming baselines in
terms of quality, diversity, and face-voice correspondence.
|
2302.14045 | Li Dong | Shaohan Huang, Li Dong, Wenhui Wang, Yaru Hao, Saksham Singhal,
Shuming Ma, Tengchao Lv, Lei Cui, Owais Khan Mohammed, Barun Patra, Qiang
Liu, Kriti Aggarwal, Zewen Chi, Johan Bjorck, Vishrav Chaudhary, Subhojit
Som, Xia Song, Furu Wei | Language Is Not All You Need: Aligning Perception with Language Models | null | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A big convergence of language, multimodal perception, action, and world
modeling is a key step toward artificial general intelligence. In this work, we
introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive
general modalities, learn in context (i.e., few-shot), and follow instructions
(i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale
multimodal corpora, including arbitrarily interleaved text and images,
image-caption pairs, and text data. We evaluate various settings, including
zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range
of tasks without any gradient updates or finetuning. Experimental results show
that Kosmos-1 achieves impressive performance on (i) language understanding,
generation, and even OCR-free NLP (directly fed with document images), (ii)
perception-language tasks, including multimodal dialogue, image captioning,
visual question answering, and (iii) vision tasks, such as image recognition
with descriptions (specifying classification via text instructions). We also
show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge
from language to multimodal, and from multimodal to language. In addition, we
introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning
capability of MLLMs.
| [
{
"version": "v1",
"created": "Mon, 27 Feb 2023 18:55:27 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 11:04:51 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Huang",
"Shaohan",
""
],
[
"Dong",
"Li",
""
],
[
"Wang",
"Wenhui",
""
],
[
"Hao",
"Yaru",
""
],
[
"Singhal",
"Saksham",
""
],
[
"Ma",
"Shuming",
""
],
[
"Lv",
"Tengchao",
""
],
[
"Cui",
"Lei",
""
],
[
"Mohammed",
"Owais Khan",
""
],
[
"Patra",
"Barun",
""
],
[
"Liu",
"Qiang",
""
],
[
"Aggarwal",
"Kriti",
""
],
[
"Chi",
"Zewen",
""
],
[
"Bjorck",
"Johan",
""
],
[
"Chaudhary",
"Vishrav",
""
],
[
"Som",
"Subhojit",
""
],
[
"Song",
"Xia",
""
],
[
"Wei",
"Furu",
""
]
] | TITLE: Language Is Not All You Need: Aligning Perception with Language Models
ABSTRACT: A big convergence of language, multimodal perception, action, and world
modeling is a key step toward artificial general intelligence. In this work, we
introduce Kosmos-1, a Multimodal Large Language Model (MLLM) that can perceive
general modalities, learn in context (i.e., few-shot), and follow instructions
(i.e., zero-shot). Specifically, we train Kosmos-1 from scratch on web-scale
multimodal corpora, including arbitrarily interleaved text and images,
image-caption pairs, and text data. We evaluate various settings, including
zero-shot, few-shot, and multimodal chain-of-thought prompting, on a wide range
of tasks without any gradient updates or finetuning. Experimental results show
that Kosmos-1 achieves impressive performance on (i) language understanding,
generation, and even OCR-free NLP (directly fed with document images), (ii)
perception-language tasks, including multimodal dialogue, image captioning,
visual question answering, and (iii) vision tasks, such as image recognition
with descriptions (specifying classification via text instructions). We also
show that MLLMs can benefit from cross-modal transfer, i.e., transfer knowledge
from language to multimodal, and from multimodal to language. In addition, we
introduce a dataset of Raven IQ test, which diagnoses the nonverbal reasoning
capability of MLLMs.
|
2302.14350 | Xianglong Lang | Xianglong Lang, Zhuming Wang, Zun Li, Meng Tian, Ge Shi, Lifang Wu and
Liang Wang | Knowledge Augmented Relation Inference for Group Activity Recognition | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most existing group activity recognition methods construct spatial-temporal
relations merely based on visual representation. Some methods introduce extra
knowledge, such as action labels, to build semantic relations and use them to
refine the visual presentation. However, the knowledge they explored just stay
at the semantic-level, which is insufficient for pursing notable accuracy. In
this paper, we propose to exploit knowledge concretization for the group
activity recognition, and develop a novel Knowledge Augmented Relation
Inference framework that can effectively use the concretized knowledge to
improve the individual representations. Specifically, the framework consists of
a Visual Representation Module to extract individual appearance features, a
Knowledge Augmented Semantic Relation Module explore semantic representations
of individual actions, and a Knowledge-Semantic-Visual Interaction Module aims
to integrate visual and semantic information by the knowledge. Benefiting from
these modules, the proposed framework can utilize knowledge to enhance the
relation inference process and the individual representations, thus improving
the performance of group activity recognition. Experimental results on two
public datasets show that the proposed framework achieves competitive
performance compared with state-of-the-art methods.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 06:59:05 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Mar 2023 08:12:08 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Lang",
"Xianglong",
""
],
[
"Wang",
"Zhuming",
""
],
[
"Li",
"Zun",
""
],
[
"Tian",
"Meng",
""
],
[
"Shi",
"Ge",
""
],
[
"Wu",
"Lifang",
""
],
[
"Wang",
"Liang",
""
]
] | TITLE: Knowledge Augmented Relation Inference for Group Activity Recognition
ABSTRACT: Most existing group activity recognition methods construct spatial-temporal
relations merely based on visual representation. Some methods introduce extra
knowledge, such as action labels, to build semantic relations and use them to
refine the visual presentation. However, the knowledge they explored just stay
at the semantic-level, which is insufficient for pursing notable accuracy. In
this paper, we propose to exploit knowledge concretization for the group
activity recognition, and develop a novel Knowledge Augmented Relation
Inference framework that can effectively use the concretized knowledge to
improve the individual representations. Specifically, the framework consists of
a Visual Representation Module to extract individual appearance features, a
Knowledge Augmented Semantic Relation Module explore semantic representations
of individual actions, and a Knowledge-Semantic-Visual Interaction Module aims
to integrate visual and semantic information by the knowledge. Benefiting from
these modules, the proposed framework can utilize knowledge to enhance the
relation inference process and the individual representations, thus improving
the performance of group activity recognition. Experimental results on two
public datasets show that the proposed framework achieves competitive
performance compared with state-of-the-art methods.
|
2303.00002 | Wenbiao Yan | Wenbiao Yan, Jihua Zhu, Yiyang Zhou, Yifei Wang, Qinghai Zheng | Multi-view Semantic Consistency based Information Bottleneck for
Clustering | 9 pages, 4 figures, conference | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-view clustering can make use of multi-source information for
unsupervised clustering. Most existing methods focus on learning a fused
representation matrix, while ignoring the influence of private information and
noise. To address this limitation, we introduce a novel Multi-view Semantic
Consistency based Information Bottleneck for clustering (MSCIB). Specifically,
MSCIB pursues semantic consistency to improve the learning process of
information bottleneck for different views. It conducts the alignment operation
of multiple views in the semantic space and jointly achieves the valuable
consistent information of multi-view data. In this way, the learned semantic
consistency from multi-view data can improve the information bottleneck to more
exactly distinguish the consistent information and learn a unified feature
representation with more discriminative consistent information for clustering.
Experiments on various types of multi-view datasets show that MSCIB achieves
state-of-the-art performance.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 02:01:58 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Yan",
"Wenbiao",
""
],
[
"Zhu",
"Jihua",
""
],
[
"Zhou",
"Yiyang",
""
],
[
"Wang",
"Yifei",
""
],
[
"Zheng",
"Qinghai",
""
]
] | TITLE: Multi-view Semantic Consistency based Information Bottleneck for
Clustering
ABSTRACT: Multi-view clustering can make use of multi-source information for
unsupervised clustering. Most existing methods focus on learning a fused
representation matrix, while ignoring the influence of private information and
noise. To address this limitation, we introduce a novel Multi-view Semantic
Consistency based Information Bottleneck for clustering (MSCIB). Specifically,
MSCIB pursues semantic consistency to improve the learning process of
information bottleneck for different views. It conducts the alignment operation
of multiple views in the semantic space and jointly achieves the valuable
consistent information of multi-view data. In this way, the learned semantic
consistency from multi-view data can improve the information bottleneck to more
exactly distinguish the consistent information and learn a unified feature
representation with more discriminative consistent information for clustering.
Experiments on various types of multi-view datasets show that MSCIB achieves
state-of-the-art performance.
|
2303.00050 | Decai Chen | Decai Chen, Haofei Lu, Ingo Feldmann, Oliver Schreer, Peter Eisert | Dynamic Multi-View Scene Reconstruction Using Neural Implicit Surface | 5 pages, accepted by ICASSP 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Reconstructing general dynamic scenes is important for many computer vision
and graphics applications. Recent works represent the dynamic scene with neural
radiance fields for photorealistic view synthesis, while their surface geometry
is under-constrained and noisy. Other works introduce surface constraints to
the implicit neural representation to disentangle the ambiguity of geometry and
appearance field for static scene reconstruction. To bridge the gap between
rendering dynamic scenes and recovering static surface geometry, we propose a
template-free method to reconstruct surface geometry and appearance using
neural implicit representations from multi-view videos. We leverage
topology-aware deformation and the signed distance field to learn complex
dynamic surfaces via differentiable volume rendering without scene-specific
prior knowledge like template models. Furthermore, we propose a novel
mask-based ray selection strategy to significantly boost the optimization on
challenging time-varying regions. Experiments on different multi-view video
datasets demonstrate that our method achieves high-fidelity surface
reconstruction as well as photorealistic novel view synthesis.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 19:47:30 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Chen",
"Decai",
""
],
[
"Lu",
"Haofei",
""
],
[
"Feldmann",
"Ingo",
""
],
[
"Schreer",
"Oliver",
""
],
[
"Eisert",
"Peter",
""
]
] | TITLE: Dynamic Multi-View Scene Reconstruction Using Neural Implicit Surface
ABSTRACT: Reconstructing general dynamic scenes is important for many computer vision
and graphics applications. Recent works represent the dynamic scene with neural
radiance fields for photorealistic view synthesis, while their surface geometry
is under-constrained and noisy. Other works introduce surface constraints to
the implicit neural representation to disentangle the ambiguity of geometry and
appearance field for static scene reconstruction. To bridge the gap between
rendering dynamic scenes and recovering static surface geometry, we propose a
template-free method to reconstruct surface geometry and appearance using
neural implicit representations from multi-view videos. We leverage
topology-aware deformation and the signed distance field to learn complex
dynamic surfaces via differentiable volume rendering without scene-specific
prior knowledge like template models. Furthermore, we propose a novel
mask-based ray selection strategy to significantly boost the optimization on
challenging time-varying regions. Experiments on different multi-view video
datasets demonstrate that our method achieves high-fidelity surface
reconstruction as well as photorealistic novel view synthesis.
|
2303.00058 | Jamie Haddock | Tyler Will, Runyu Zhang, Eli Sadovnik, Mengdi Gao, Joshua Vendrow,
Jamie Haddock, Denali Molitor, Deanna Needell | Neural Nonnegative Matrix Factorization for Hierarchical Multilayer
Topic Modeling | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a new method based on nonnegative matrix factorization, Neural
NMF, for detecting latent hierarchical structure in data. Datasets with
hierarchical structure arise in a wide variety of fields, such as document
classification, image processing, and bioinformatics. Neural NMF recursively
applies NMF in layers to discover overarching topics encompassing the
lower-level features. We derive a backpropagation optimization scheme that
allows us to frame hierarchical NMF as a neural network. We test Neural NMF on
a synthetic hierarchical dataset, the 20 Newsgroups dataset, and the MyLymeData
symptoms dataset. Numerical results demonstrate that Neural NMF outperforms
other hierarchical NMF methods on these data sets and offers better learned
hierarchical structure and interpretability of topics.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 20:00:16 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Will",
"Tyler",
""
],
[
"Zhang",
"Runyu",
""
],
[
"Sadovnik",
"Eli",
""
],
[
"Gao",
"Mengdi",
""
],
[
"Vendrow",
"Joshua",
""
],
[
"Haddock",
"Jamie",
""
],
[
"Molitor",
"Denali",
""
],
[
"Needell",
"Deanna",
""
]
] | TITLE: Neural Nonnegative Matrix Factorization for Hierarchical Multilayer
Topic Modeling
ABSTRACT: We introduce a new method based on nonnegative matrix factorization, Neural
NMF, for detecting latent hierarchical structure in data. Datasets with
hierarchical structure arise in a wide variety of fields, such as document
classification, image processing, and bioinformatics. Neural NMF recursively
applies NMF in layers to discover overarching topics encompassing the
lower-level features. We derive a backpropagation optimization scheme that
allows us to frame hierarchical NMF as a neural network. We test Neural NMF on
a synthetic hierarchical dataset, the 20 Newsgroups dataset, and the MyLymeData
symptoms dataset. Numerical results demonstrate that Neural NMF outperforms
other hierarchical NMF methods on these data sets and offers better learned
hierarchical structure and interpretability of topics.
|
2303.00091 | Jong Chul Ye | Jaeyoung Huh, Sangjoon Park, Jeong Eun Lee, Jong Chul Ye | Improving Medical Speech-to-Text Accuracy with Vision-Language
Pre-training Model | null | null | null | null | eess.AS cs.AI cs.CL cs.CV cs.SD eess.IV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Automatic Speech Recognition (ASR) is a technology that converts spoken words
into text, facilitating interaction between humans and machines. One of the
most common applications of ASR is Speech-To-Text (STT) technology, which
simplifies user workflows by transcribing spoken words into text. In the
medical field, STT has the potential to significantly reduce the workload of
clinicians who rely on typists to transcribe their voice recordings. However,
developing an STT model for the medical domain is challenging due to the lack
of sufficient speech and text datasets. To address this issue, we propose a
medical-domain text correction method that modifies the output text of a
general STT system using the Vision Language Pre-training (VLP) method. VLP
combines textual and visual information to correct text based on image
knowledge. Our extensive experiments demonstrate that the proposed method
offers quantitatively and clinically significant improvements in STT
performance in the medical field. We further show that multi-modal
understanding of image and text information outperforms single-modal
understanding using only text information.
| [
{
"version": "v1",
"created": "Mon, 27 Feb 2023 08:06:04 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Huh",
"Jaeyoung",
""
],
[
"Park",
"Sangjoon",
""
],
[
"Lee",
"Jeong Eun",
""
],
[
"Ye",
"Jong Chul",
""
]
] | TITLE: Improving Medical Speech-to-Text Accuracy with Vision-Language
Pre-training Model
ABSTRACT: Automatic Speech Recognition (ASR) is a technology that converts spoken words
into text, facilitating interaction between humans and machines. One of the
most common applications of ASR is Speech-To-Text (STT) technology, which
simplifies user workflows by transcribing spoken words into text. In the
medical field, STT has the potential to significantly reduce the workload of
clinicians who rely on typists to transcribe their voice recordings. However,
developing an STT model for the medical domain is challenging due to the lack
of sufficient speech and text datasets. To address this issue, we propose a
medical-domain text correction method that modifies the output text of a
general STT system using the Vision Language Pre-training (VLP) method. VLP
combines textual and visual information to correct text based on image
knowledge. Our extensive experiments demonstrate that the proposed method
offers quantitatively and clinically significant improvements in STT
performance in the medical field. We further show that multi-modal
understanding of image and text information outperforms single-modal
understanding using only text information.
|
2303.00106 | Zhenmei Shi | Zhenmei Shi, Jiefeng Chen, Kunyang Li, Jayaram Raghuram, Xi Wu, Yingyu
Liang, Somesh Jha | The Trade-off between Universality and Label Efficiency of
Representations from Contrastive Learning | 42 pages | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pre-training representations (a.k.a. foundation models) has recently become a
prevalent learning paradigm, where one first pre-trains a representation using
large-scale unlabeled data, and then learns simple predictors on top of the
representation using small labeled data from the downstream tasks. There are
two key desiderata for the representation: label efficiency (the ability to
learn an accurate classifier on top of the representation with a small amount
of labeled data) and universality (usefulness across a wide range of downstream
tasks). In this paper, we focus on one of the most popular instantiations of
this paradigm: contrastive learning with linear probing, i.e., learning a
linear predictor on the representation pre-trained by contrastive learning. We
show that there exists a trade-off between the two desiderata so that one may
not be able to achieve both simultaneously. Specifically, we provide analysis
using a theoretical data model and show that, while more diverse pre-training
data result in more diverse features for different tasks (improving
universality), it puts less emphasis on task-specific features, giving rise to
larger sample complexity for down-stream supervised tasks, and thus worse
prediction performance. Guided by this analysis, we propose a contrastive
regularization method to improve the trade-off. We validate our analysis and
method empirically with systematic experiments using real-world datasets and
foundation models.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 22:14:33 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Shi",
"Zhenmei",
""
],
[
"Chen",
"Jiefeng",
""
],
[
"Li",
"Kunyang",
""
],
[
"Raghuram",
"Jayaram",
""
],
[
"Wu",
"Xi",
""
],
[
"Liang",
"Yingyu",
""
],
[
"Jha",
"Somesh",
""
]
] | TITLE: The Trade-off between Universality and Label Efficiency of
Representations from Contrastive Learning
ABSTRACT: Pre-training representations (a.k.a. foundation models) has recently become a
prevalent learning paradigm, where one first pre-trains a representation using
large-scale unlabeled data, and then learns simple predictors on top of the
representation using small labeled data from the downstream tasks. There are
two key desiderata for the representation: label efficiency (the ability to
learn an accurate classifier on top of the representation with a small amount
of labeled data) and universality (usefulness across a wide range of downstream
tasks). In this paper, we focus on one of the most popular instantiations of
this paradigm: contrastive learning with linear probing, i.e., learning a
linear predictor on the representation pre-trained by contrastive learning. We
show that there exists a trade-off between the two desiderata so that one may
not be able to achieve both simultaneously. Specifically, we provide analysis
using a theoretical data model and show that, while more diverse pre-training
data result in more diverse features for different tasks (improving
universality), it puts less emphasis on task-specific features, giving rise to
larger sample complexity for down-stream supervised tasks, and thus worse
prediction performance. Guided by this analysis, we propose a contrastive
regularization method to improve the trade-off. We validate our analysis and
method empirically with systematic experiments using real-world datasets and
foundation models.
|
2303.00135 | Rohitash Chandra | Janhavi Lande, Arti Pillay, Rohitash Chandra | Deep learning for COVID-19 topic modelling via Twitter: Alpha, Delta and
Omicron | null | null | null | null | cs.LG cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | Topic modelling with innovative deep learning methods has gained interest for
a wide range of applications that includes COVID-19. Topic modelling can
provide, psychological, social and cultural insights for understanding human
behaviour in extreme events such as the COVID-19 pandemic. In this paper, we
use prominent deep learning-based language models for COVID-19 topic modelling
taking into account data from emergence (Alpha) to the Omicron variant. We
apply topic modeling to review the public behaviour across the first, second
and third waves based on Twitter dataset from India. Our results show that the
topics extracted for the subsequent waves had certain overlapping themes such
as covers governance, vaccination, and pandemic management while novel issues
aroused in political, social and economic situation during COVID-19 pandemic.
We also found a strong correlation of the major topics qualitatively to news
media prevalent at the respective time period. Hence, our framework has the
potential to capture major issues arising during different phases of the
COVID-19 pandemic which can be extended to other countries and regions.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 23:40:41 GMT"
}
] | 2023-03-02T00:00:00 | [
[
"Lande",
"Janhavi",
""
],
[
"Pillay",
"Arti",
""
],
[
"Chandra",
"Rohitash",
""
]
] | TITLE: Deep learning for COVID-19 topic modelling via Twitter: Alpha, Delta and
Omicron
ABSTRACT: Topic modelling with innovative deep learning methods has gained interest for
a wide range of applications that includes COVID-19. Topic modelling can
provide, psychological, social and cultural insights for understanding human
behaviour in extreme events such as the COVID-19 pandemic. In this paper, we
use prominent deep learning-based language models for COVID-19 topic modelling
taking into account data from emergence (Alpha) to the Omicron variant. We
apply topic modeling to review the public behaviour across the first, second
and third waves based on Twitter dataset from India. Our results show that the
topics extracted for the subsequent waves had certain overlapping themes such
as covers governance, vaccination, and pandemic management while novel issues
aroused in political, social and economic situation during COVID-19 pandemic.
We also found a strong correlation of the major topics qualitatively to news
media prevalent at the respective time period. Hence, our framework has the
potential to capture major issues arising during different phases of the
COVID-19 pandemic which can be extended to other countries and regions.
|
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.