Dataset Viewer
Search is not available for this dataset
id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence | prompt
string |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2204.07661 | Soumyajit Gupta | Soumyajit Gupta, Venelin Kovatchev, Anubrata Das, Maria De-Arteaga,
Matthew Lease | Finding Pareto Trade-offs in Fair and Accurate Detection of Toxic Speech | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Optimizing NLP models for fairness poses many challenges. Lack of
differentiable fairness measures prevents gradient-based loss training or
requires surrogate losses that diverge from the true metric of interest. In
addition, competing objectives (e.g., accuracy vs. fairness) often require
making trade-offs based on stakeholder preferences, but stakeholders may not
know their preferences before seeing system performance under different
trade-off settings. To address these challenges, we begin by formulating a
differentiable version of a popular fairness measure, Accuracy Parity, to
provide balanced accuracy across demographic groups. Next, we show how
model-agnostic, HyperNetwork optimization can efficiently train arbitrary NLP
model architectures to learn Pareto-optimal trade-offs between competing
metrics. Focusing on the task of toxic language detection, we show the
generality and efficacy of our methods across two datasets, three neural
architectures, and three fairness losses.
| [
{
"version": "v1",
"created": "Fri, 15 Apr 2022 22:11:25 GMT"
},
{
"version": "v2",
"created": "Tue, 10 May 2022 18:36:41 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 00:29:44 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Gupta",
"Soumyajit",
""
],
[
"Kovatchev",
"Venelin",
""
],
[
"Das",
"Anubrata",
""
],
[
"De-Arteaga",
"Maria",
""
],
[
"Lease",
"Matthew",
""
]
] | TITLE: Finding Pareto Trade-offs in Fair and Accurate Detection of Toxic Speech
ABSTRACT: Optimizing NLP models for fairness poses many challenges. Lack of
differentiable fairness measures prevents gradient-based loss training or
requires surrogate losses that diverge from the true metric of interest. In
addition, competing objectives (e.g., accuracy vs. fairness) often require
making trade-offs based on stakeholder preferences, but stakeholders may not
know their preferences before seeing system performance under different
trade-off settings. To address these challenges, we begin by formulating a
differentiable version of a popular fairness measure, Accuracy Parity, to
provide balanced accuracy across demographic groups. Next, we show how
model-agnostic, HyperNetwork optimization can efficiently train arbitrary NLP
model architectures to learn Pareto-optimal trade-offs between competing
metrics. Focusing on the task of toxic language detection, we show the
generality and efficacy of our methods across two datasets, three neural
architectures, and three fairness losses.
|
2303.17408 | Yucheng Ruan | Yucheng Ruan, Xiang Lan, Daniel J. Tan, Hairil Rizal Abdullah,
Mengling Feng | P-Transformer: A Prompt-based Multimodal Transformer Architecture For
Medical Tabular Data | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Medical tabular data, abundant in Electronic Health Records (EHRs), is a
valuable resource for diverse medical tasks such as risk prediction. While deep
learning approaches, particularly transformer-based models, have shown
remarkable performance in tabular data prediction, there are still problems
remaining for existing work to be effectively adapted into medical domain, such
as ignoring unstructured free-texts and underutilizing the textual information
in structured data. To address these issues, we propose PTransformer, a
\underline{P}rompt-based multimodal \underline{Transformer} architecture
designed specifically for medical tabular data. This framework consists of two
critical components: a tabular cell embedding generator and a tabular
transformer. The former efficiently encodes diverse modalities from both
structured and unstructured tabular data into a harmonized language semantic
space with the help of pre-trained sentence encoder and medical prompts. The
latter integrates cell representations to generate patient embeddings for
various medical tasks. In comprehensive experiments on two real-world datasets
for three medical tasks, PTransformer demonstrated the improvements with
10.9%/11.0% on RMSE/MAE, 0.5%/2.2% on RMSE/MAE, and 1.6%/0.8% on BACC/AUROC
compared to state-of-the-art (SOTA) baselines in predictability.
| [
{
"version": "v1",
"created": "Thu, 30 Mar 2023 14:25:44 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Aug 2023 08:58:25 GMT"
},
{
"version": "v3",
"created": "Tue, 9 Jan 2024 10:28:00 GMT"
},
{
"version": "v4",
"created": "Thu, 10 Apr 2025 06:24:36 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Ruan",
"Yucheng",
""
],
[
"Lan",
"Xiang",
""
],
[
"Tan",
"Daniel J.",
""
],
[
"Abdullah",
"Hairil Rizal",
""
],
[
"Feng",
"Mengling",
""
]
] | TITLE: P-Transformer: A Prompt-based Multimodal Transformer Architecture For
Medical Tabular Data
ABSTRACT: Medical tabular data, abundant in Electronic Health Records (EHRs), is a
valuable resource for diverse medical tasks such as risk prediction. While deep
learning approaches, particularly transformer-based models, have shown
remarkable performance in tabular data prediction, there are still problems
remaining for existing work to be effectively adapted into medical domain, such
as ignoring unstructured free-texts and underutilizing the textual information
in structured data. To address these issues, we propose PTransformer, a
\underline{P}rompt-based multimodal \underline{Transformer} architecture
designed specifically for medical tabular data. This framework consists of two
critical components: a tabular cell embedding generator and a tabular
transformer. The former efficiently encodes diverse modalities from both
structured and unstructured tabular data into a harmonized language semantic
space with the help of pre-trained sentence encoder and medical prompts. The
latter integrates cell representations to generate patient embeddings for
various medical tasks. In comprehensive experiments on two real-world datasets
for three medical tasks, PTransformer demonstrated the improvements with
10.9%/11.0% on RMSE/MAE, 0.5%/2.2% on RMSE/MAE, and 1.6%/0.8% on BACC/AUROC
compared to state-of-the-art (SOTA) baselines in predictability.
|
2305.03535 | Jonas Hein | Jonas Hein, Nicola Cavalcanti, Daniel Suter, Lukas Zingg, Fabio
Carrillo, Lilian Calvet, Mazda Farshad, Marc Pollefeys, Nassir Navab, Philipp
F\"urnstahl | Next-generation Surgical Navigation: Marker-less Multi-view 6DoF Pose
Estimation of Surgical Instruments | Accepted for publication in Medical Image Analysis. Project page:
https://jonashein.github.io/mvpsp/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | State-of-the-art research of traditional computer vision is increasingly
leveraged in the surgical domain. A particular focus in computer-assisted
surgery is to replace marker-based tracking systems for instrument localization
with pure image-based 6DoF pose estimation using deep-learning methods.
However, state-of-the-art single-view pose estimation methods do not yet meet
the accuracy required for surgical navigation. In this context, we investigate
the benefits of multi-view setups for highly accurate and occlusion-robust 6DoF
pose estimation of surgical instruments and derive recommendations for an ideal
camera system that addresses the challenges in the operating room.
The contributions of this work are threefold. First, we present a
multi-camera capture setup consisting of static and head-mounted cameras, which
allows us to study the performance of pose estimation methods under various
camera configurations. Second, we publish a multi-view RGB-D video dataset of
ex-vivo spine surgeries, captured in a surgical wet lab and a real operating
theatre and including rich annotations for surgeon, instrument, and patient
anatomy. Third, we evaluate three state-of-the-art single-view and multi-view
methods for the task of 6DoF pose estimation of surgical instruments and
analyze the influence of camera configurations, training data, and occlusions
on the pose accuracy and generalization ability. The best method utilizes five
cameras in a multi-view pose optimization and achieves an average position and
orientation error of 1.01 mm and 0.89{\deg} for a surgical drill as well as
2.79 mm and 3.33{\deg} for a screwdriver under optimal conditions. Our results
demonstrate that marker-less tracking of surgical instruments is becoming a
feasible alternative to existing marker-based systems.
| [
{
"version": "v1",
"created": "Fri, 5 May 2023 13:42:19 GMT"
},
{
"version": "v2",
"created": "Fri, 22 Dec 2023 20:52:50 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 17:23:33 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Hein",
"Jonas",
""
],
[
"Cavalcanti",
"Nicola",
""
],
[
"Suter",
"Daniel",
""
],
[
"Zingg",
"Lukas",
""
],
[
"Carrillo",
"Fabio",
""
],
[
"Calvet",
"Lilian",
""
],
[
"Farshad",
"Mazda",
""
],
[
"Pollefeys",
"Marc",
""
],
[
"Navab",
"Nassir",
""
],
[
"Fürnstahl",
"Philipp",
""
]
] | TITLE: Next-generation Surgical Navigation: Marker-less Multi-view 6DoF Pose
Estimation of Surgical Instruments
ABSTRACT: State-of-the-art research of traditional computer vision is increasingly
leveraged in the surgical domain. A particular focus in computer-assisted
surgery is to replace marker-based tracking systems for instrument localization
with pure image-based 6DoF pose estimation using deep-learning methods.
However, state-of-the-art single-view pose estimation methods do not yet meet
the accuracy required for surgical navigation. In this context, we investigate
the benefits of multi-view setups for highly accurate and occlusion-robust 6DoF
pose estimation of surgical instruments and derive recommendations for an ideal
camera system that addresses the challenges in the operating room.
The contributions of this work are threefold. First, we present a
multi-camera capture setup consisting of static and head-mounted cameras, which
allows us to study the performance of pose estimation methods under various
camera configurations. Second, we publish a multi-view RGB-D video dataset of
ex-vivo spine surgeries, captured in a surgical wet lab and a real operating
theatre and including rich annotations for surgeon, instrument, and patient
anatomy. Third, we evaluate three state-of-the-art single-view and multi-view
methods for the task of 6DoF pose estimation of surgical instruments and
analyze the influence of camera configurations, training data, and occlusions
on the pose accuracy and generalization ability. The best method utilizes five
cameras in a multi-view pose optimization and achieves an average position and
orientation error of 1.01 mm and 0.89{\deg} for a surgical drill as well as
2.79 mm and 3.33{\deg} for a screwdriver under optimal conditions. Our results
demonstrate that marker-less tracking of surgical instruments is becoming a
feasible alternative to existing marker-based systems.
|
2305.15932 | Jie He | Jie He and Simon Chi Lok U and V\'ictor Guti\'errez-Basulto and Jeff
Z. Pan | BUCA: A Binary Classification Approach to Unsupervised Commonsense
Question Answering | There is a text error in Table 10 | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Unsupervised commonsense reasoning (UCR) is becoming increasingly popular as
the construction of commonsense reasoning datasets is expensive, and they are
inevitably limited in their scope. A popular approach to UCR is to fine-tune
language models with external knowledge (e.g., knowledge graphs), but this
usually requires a large number of training examples. In this paper, we propose
to transform the downstream multiple choice question answering task into a
simpler binary classification task by ranking all candidate answers according
to their reasonableness. To this end, for training the model, we convert the
knowledge graph triples into reasonable and unreasonable texts. Extensive
experimental results show the effectiveness of our approach on various multiple
choice question answering benchmarks. Furthermore, compared with existing UCR
approaches using KGs, ours is less data hungry. Our code is available at
https://github.com/probe2/BUCA.
| [
{
"version": "v1",
"created": "Thu, 25 May 2023 10:59:47 GMT"
},
{
"version": "v2",
"created": "Wed, 7 Jun 2023 20:33:09 GMT"
},
{
"version": "v3",
"created": "Mon, 10 Mar 2025 09:28:29 GMT"
},
{
"version": "v4",
"created": "Wed, 9 Apr 2025 21:40:10 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"He",
"Jie",
""
],
[
"U",
"Simon Chi Lok",
""
],
[
"Gutiérrez-Basulto",
"Víctor",
""
],
[
"Pan",
"Jeff Z.",
""
]
] | TITLE: BUCA: A Binary Classification Approach to Unsupervised Commonsense
Question Answering
ABSTRACT: Unsupervised commonsense reasoning (UCR) is becoming increasingly popular as
the construction of commonsense reasoning datasets is expensive, and they are
inevitably limited in their scope. A popular approach to UCR is to fine-tune
language models with external knowledge (e.g., knowledge graphs), but this
usually requires a large number of training examples. In this paper, we propose
to transform the downstream multiple choice question answering task into a
simpler binary classification task by ranking all candidate answers according
to their reasonableness. To this end, for training the model, we convert the
knowledge graph triples into reasonable and unreasonable texts. Extensive
experimental results show the effectiveness of our approach on various multiple
choice question answering benchmarks. Furthermore, compared with existing UCR
approaches using KGs, ours is less data hungry. Our code is available at
https://github.com/probe2/BUCA.
|
2308.07223 | Melanie Roschewitz | M\'elanie Roschewitz and Ben Glocker | Distance Matters For Improving Performance Estimation Under Covariate
Shift | Accepted to ICCV Workshop on Uncertainty Quantification for Computer
Vision 2023 | Proceedings of the IEEE/CVF International Conference on Computer
Vision (ICCV) Workshops, 2023, pp. 4549-4559 | 10.1109/ICCVW60793.2023.00489 | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Performance estimation under covariate shift is a crucial component of safe
AI model deployment, especially for sensitive use-cases. Recently, several
solutions were proposed to tackle this problem, most leveraging model
predictions or softmax confidence to derive accuracy estimates. However, under
dataset shifts, confidence scores may become ill-calibrated if samples are too
far from the training distribution. In this work, we show that taking into
account distances of test samples to their expected training distribution can
significantly improve performance estimation under covariate shift. Precisely,
we introduce a "distance-check" to flag samples that lie too far from the
expected distribution, to avoid relying on their untrustworthy model outputs in
the accuracy estimation step. We demonstrate the effectiveness of this method
on 13 image classification tasks, across a wide-range of natural and synthetic
distribution shifts and hundreds of models, with a median relative MAE
improvement of 27% over the best baseline across all tasks, and SOTA
performance on 10 out of 13 tasks. Our code is publicly available at
https://github.com/melanibe/distance_matters_performance_estimation.
| [
{
"version": "v1",
"created": "Mon, 14 Aug 2023 15:49:19 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Roschewitz",
"Mélanie",
""
],
[
"Glocker",
"Ben",
""
]
] | TITLE: Distance Matters For Improving Performance Estimation Under Covariate
Shift
ABSTRACT: Performance estimation under covariate shift is a crucial component of safe
AI model deployment, especially for sensitive use-cases. Recently, several
solutions were proposed to tackle this problem, most leveraging model
predictions or softmax confidence to derive accuracy estimates. However, under
dataset shifts, confidence scores may become ill-calibrated if samples are too
far from the training distribution. In this work, we show that taking into
account distances of test samples to their expected training distribution can
significantly improve performance estimation under covariate shift. Precisely,
we introduce a "distance-check" to flag samples that lie too far from the
expected distribution, to avoid relying on their untrustworthy model outputs in
the accuracy estimation step. We demonstrate the effectiveness of this method
on 13 image classification tasks, across a wide-range of natural and synthetic
distribution shifts and hundreds of models, with a median relative MAE
improvement of 27% over the best baseline across all tasks, and SOTA
performance on 10 out of 13 tasks. Our code is publicly available at
https://github.com/melanibe/distance_matters_performance_estimation.
|
2309.15408 | Mario Beraha | Mario Beraha, Stefano Favaro, Matteo Sesia | A smoothed-Bayesian approach to frequency recovery from sketched data | null | null | null | null | stat.ME cs.DS cs.IR math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We provide a novel statistical perspective on a classical problem at the
intersection of computer science and information theory: recovering the
empirical frequency of a symbol in a large discrete dataset using only a
compressed representation, or sketch, obtained via random hashing. Departing
from traditional algorithmic approaches, recent works have proposed Bayesian
nonparametric (BNP) methods that can provide more informative frequency
estimates by leveraging modeling assumptions about the distribution of the
sketched data. In this paper, we propose a smoothed-Bayesian method, inspired
by existing BNP approaches but designed in a frequentist framework to overcome
the computational limitations of the BNP approaches when dealing with
large-scale data from realistic distributions, including those with power-law
tail behaviors. For sketches obtained with a single hash function, our approach
is supported by rigorous frequentist properties, including unbiasedness and
optimality under a squared error loss function within an intuitive class of
linear estimators. For sketches with multiple hash functions, we introduce an
approach based on multi-view learning to construct computationally efficient
frequency estimators. We validate our method on synthetic and real data,
comparing its performance to that of existing alternatives.
| [
{
"version": "v1",
"created": "Wed, 27 Sep 2023 05:20:53 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Jun 2024 13:15:11 GMT"
},
{
"version": "v3",
"created": "Thu, 28 Nov 2024 13:46:38 GMT"
},
{
"version": "v4",
"created": "Thu, 10 Apr 2025 08:21:29 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Beraha",
"Mario",
""
],
[
"Favaro",
"Stefano",
""
],
[
"Sesia",
"Matteo",
""
]
] | TITLE: A smoothed-Bayesian approach to frequency recovery from sketched data
ABSTRACT: We provide a novel statistical perspective on a classical problem at the
intersection of computer science and information theory: recovering the
empirical frequency of a symbol in a large discrete dataset using only a
compressed representation, or sketch, obtained via random hashing. Departing
from traditional algorithmic approaches, recent works have proposed Bayesian
nonparametric (BNP) methods that can provide more informative frequency
estimates by leveraging modeling assumptions about the distribution of the
sketched data. In this paper, we propose a smoothed-Bayesian method, inspired
by existing BNP approaches but designed in a frequentist framework to overcome
the computational limitations of the BNP approaches when dealing with
large-scale data from realistic distributions, including those with power-law
tail behaviors. For sketches obtained with a single hash function, our approach
is supported by rigorous frequentist properties, including unbiasedness and
optimality under a squared error loss function within an intuitive class of
linear estimators. For sketches with multiple hash functions, we introduce an
approach based on multi-view learning to construct computationally efficient
frequency estimators. We validate our method on synthetic and real data,
comparing its performance to that of existing alternatives.
|
2311.06293 | Zeynab Kaseb | Zeynab Kaseb, Matthias Moller, Giorgio Tosti Balducci, Peter Palensky,
Pedro P. Vergara | Quantum Neural Networks for Power Flow Analysis | 8 pages, 13 figures | null | 10.1016/j.epsr.2024.110677 | null | quant-ph cs.LG cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | This paper explores the potential application of quantum and hybrid
quantum-classical neural networks in power flow analysis. Experiments are
conducted using two datasets based on 4-bus and 33-bus test systems. A
systematic performance comparison is also conducted among quantum, hybrid
quantum-classical, and classical neural networks. The comparison is based on
(i) generalization ability, (ii) robustness, (iii) training dataset size
needed, (iv) training error, and (v) training process stability. The results
show that the developed hybrid quantum-classical neural network outperforms
both quantum and classical neural networks, and hence can improve deep
learning-based power flow analysis in the noisy-intermediate-scale quantum
(NISQ) and fault-tolerant quantum (FTQ) era.
| [
{
"version": "v1",
"created": "Sat, 4 Nov 2023 11:25:31 GMT"
},
{
"version": "v2",
"created": "Sun, 10 Mar 2024 15:49:57 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Kaseb",
"Zeynab",
""
],
[
"Moller",
"Matthias",
""
],
[
"Balducci",
"Giorgio Tosti",
""
],
[
"Palensky",
"Peter",
""
],
[
"Vergara",
"Pedro P.",
""
]
] | TITLE: Quantum Neural Networks for Power Flow Analysis
ABSTRACT: This paper explores the potential application of quantum and hybrid
quantum-classical neural networks in power flow analysis. Experiments are
conducted using two datasets based on 4-bus and 33-bus test systems. A
systematic performance comparison is also conducted among quantum, hybrid
quantum-classical, and classical neural networks. The comparison is based on
(i) generalization ability, (ii) robustness, (iii) training dataset size
needed, (iv) training error, and (v) training process stability. The results
show that the developed hybrid quantum-classical neural network outperforms
both quantum and classical neural networks, and hence can improve deep
learning-based power flow analysis in the noisy-intermediate-scale quantum
(NISQ) and fault-tolerant quantum (FTQ) era.
|
2311.13706 | Nicol\'as Gaggion Ph.D. | Nicol\'as Gaggion, Benjamin A. Matheson, Yan Xia, Rodrigo Bonazzola,
Nishant Ravikumar, Zeike A. Taylor, Diego H. Milone, Alejandro F. Frangi,
Enzo Ferrante | Multi-view Hybrid Graph Convolutional Network for Volume-to-mesh
Reconstruction in Cardiovascular MRI | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Cardiovascular magnetic resonance imaging is emerging as a crucial tool to
examine cardiac morphology and function. Essential to this endeavour are
anatomical 3D surface and volumetric meshes derived from CMR images, which
facilitate computational anatomy studies, biomarker discovery, and in-silico
simulations. Traditional approaches typically follow complex multi-step
pipelines, first segmenting images and then reconstructing meshes, making them
time-consuming and prone to error propagation. In response, we introduce
HybridVNet, a novel architecture for direct image-to-mesh extraction seamlessly
integrating standard convolutional neural networks with graph convolutions,
which we prove can efficiently handle surface and volumetric meshes by encoding
them as graph structures. To further enhance accuracy, we propose a multi-view
HybridVNet architecture which processes both long axis and short axis CMR,
showing that it can increase the performance of cardiac MR mesh generation. Our
model combines traditional convolutional networks with variational graph
generative models, deep supervision and mesh-specific regularisation.
Experiments on a comprehensive dataset from the UK Biobank confirm the
potential of HybridVNet to significantly advance cardiac imaging and
computational cardiology by efficiently generating high-fidelity meshes from
CMR images. Multi-view HybridVNet outperforms the state-of-the-art, achieving
improvements of up to $\sim$27\% reduction in Mean Contour Distance (from 1.86
mm to 1.35 mm for the LV Myocardium), up to $\sim$18\% improvement in Hausdorff
distance (from 4.74 mm to 3.89mm, for the LV Endocardium), and up to $\sim$8\%
in Dice Coefficient (from 0.78 to 0.84, for the LV Myocardium), highlighting
its superior accuracy.
| [
{
"version": "v1",
"created": "Wed, 22 Nov 2023 21:51:29 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Aug 2024 19:18:41 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 16:25:45 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Gaggion",
"Nicolás",
""
],
[
"Matheson",
"Benjamin A.",
""
],
[
"Xia",
"Yan",
""
],
[
"Bonazzola",
"Rodrigo",
""
],
[
"Ravikumar",
"Nishant",
""
],
[
"Taylor",
"Zeike A.",
""
],
[
"Milone",
"Diego H.",
""
],
[
"Frangi",
"Alejandro F.",
""
],
[
"Ferrante",
"Enzo",
""
]
] | TITLE: Multi-view Hybrid Graph Convolutional Network for Volume-to-mesh
Reconstruction in Cardiovascular MRI
ABSTRACT: Cardiovascular magnetic resonance imaging is emerging as a crucial tool to
examine cardiac morphology and function. Essential to this endeavour are
anatomical 3D surface and volumetric meshes derived from CMR images, which
facilitate computational anatomy studies, biomarker discovery, and in-silico
simulations. Traditional approaches typically follow complex multi-step
pipelines, first segmenting images and then reconstructing meshes, making them
time-consuming and prone to error propagation. In response, we introduce
HybridVNet, a novel architecture for direct image-to-mesh extraction seamlessly
integrating standard convolutional neural networks with graph convolutions,
which we prove can efficiently handle surface and volumetric meshes by encoding
them as graph structures. To further enhance accuracy, we propose a multi-view
HybridVNet architecture which processes both long axis and short axis CMR,
showing that it can increase the performance of cardiac MR mesh generation. Our
model combines traditional convolutional networks with variational graph
generative models, deep supervision and mesh-specific regularisation.
Experiments on a comprehensive dataset from the UK Biobank confirm the
potential of HybridVNet to significantly advance cardiac imaging and
computational cardiology by efficiently generating high-fidelity meshes from
CMR images. Multi-view HybridVNet outperforms the state-of-the-art, achieving
improvements of up to $\sim$27\% reduction in Mean Contour Distance (from 1.86
mm to 1.35 mm for the LV Myocardium), up to $\sim$18\% improvement in Hausdorff
distance (from 4.74 mm to 3.89mm, for the LV Endocardium), and up to $\sim$8\%
in Dice Coefficient (from 0.78 to 0.84, for the LV Myocardium), highlighting
its superior accuracy.
|
2401.03048 | Xin Ma | Xin Ma, Yaohui Wang, Xinyuan Chen, Gengyun Jia, Ziwei Liu, Yuan-Fang
Li, Cunjian Chen, Yu Qiao | Latte: Latent Diffusion Transformer for Video Generation | Accepted by Transactions on Machine Learning Research 2025; Project
page: https://maxin-cn.github.io/latte_project | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a novel Latent Diffusion Transformer, namely Latte, for video
generation. Latte first extracts spatio-temporal tokens from input videos and
then adopts a series of Transformer blocks to model video distribution in the
latent space. In order to model a substantial number of tokens extracted from
videos, four efficient variants are introduced from the perspective of
decomposing the spatial and temporal dimensions of input videos. To improve the
quality of generated videos, we determine the best practices of Latte through
rigorous experimental analysis, including video clip patch embedding, model
variants, timestep-class information injection, temporal positional embedding,
and learning strategies. Our comprehensive evaluation demonstrates that Latte
achieves state-of-the-art performance across four standard video generation
datasets, i.e., FaceForensics, SkyTimelapse, UCF101, and Taichi-HD. In
addition, we extend Latte to text-to-video generation (T2V) task, where Latte
achieves comparable results compared to recent T2V models. We strongly believe
that Latte provides valuable insights for future research on incorporating
Transformers into diffusion models for video generation.
| [
{
"version": "v1",
"created": "Fri, 5 Jan 2024 19:55:15 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 09:28:20 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Ma",
"Xin",
""
],
[
"Wang",
"Yaohui",
""
],
[
"Chen",
"Xinyuan",
""
],
[
"Jia",
"Gengyun",
""
],
[
"Liu",
"Ziwei",
""
],
[
"Li",
"Yuan-Fang",
""
],
[
"Chen",
"Cunjian",
""
],
[
"Qiao",
"Yu",
""
]
] | TITLE: Latte: Latent Diffusion Transformer for Video Generation
ABSTRACT: We propose a novel Latent Diffusion Transformer, namely Latte, for video
generation. Latte first extracts spatio-temporal tokens from input videos and
then adopts a series of Transformer blocks to model video distribution in the
latent space. In order to model a substantial number of tokens extracted from
videos, four efficient variants are introduced from the perspective of
decomposing the spatial and temporal dimensions of input videos. To improve the
quality of generated videos, we determine the best practices of Latte through
rigorous experimental analysis, including video clip patch embedding, model
variants, timestep-class information injection, temporal positional embedding,
and learning strategies. Our comprehensive evaluation demonstrates that Latte
achieves state-of-the-art performance across four standard video generation
datasets, i.e., FaceForensics, SkyTimelapse, UCF101, and Taichi-HD. In
addition, we extend Latte to text-to-video generation (T2V) task, where Latte
achieves comparable results compared to recent T2V models. We strongly believe
that Latte provides valuable insights for future research on incorporating
Transformers into diffusion models for video generation.
|
2402.18206 | Abhijnya Bhat | Rishubh Parihar, Abhijnya Bhat, Abhipsa Basu, Saswat Mallick, Jogendra
Nath Kundu, R. Venkatesh Babu | Balancing Act: Distribution-Guided Debiasing in Diffusion Models | CVPR 2024. Project Page : https://ab-34.github.io/balancing_act/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Diffusion Models (DMs) have emerged as powerful generative models with
unprecedented image generation capability. These models are widely used for
data augmentation and creative applications. However, DMs reflect the biases
present in the training datasets. This is especially concerning in the context
of faces, where the DM prefers one demographic subgroup vs others (eg. female
vs male). In this work, we present a method for debiasing DMs without relying
on additional data or model retraining. Specifically, we propose Distribution
Guidance, which enforces the generated images to follow the prescribed
attribute distribution. To realize this, we build on the key insight that the
latent features of denoising UNet hold rich demographic semantics, and the same
can be leveraged to guide debiased generation. We train Attribute Distribution
Predictor (ADP) - a small mlp that maps the latent features to the distribution
of attributes. ADP is trained with pseudo labels generated from existing
attribute classifiers. The proposed Distribution Guidance with ADP enables us
to do fair generation. Our method reduces bias across single/multiple
attributes and outperforms the baseline by a significant margin for
unconditional and text-conditional diffusion models. Further, we present a
downstream task of training a fair attribute classifier by rebalancing the
training set with our generated data.
| [
{
"version": "v1",
"created": "Wed, 28 Feb 2024 09:53:17 GMT"
},
{
"version": "v2",
"created": "Wed, 22 May 2024 17:23:22 GMT"
},
{
"version": "v3",
"created": "Wed, 29 May 2024 13:33:57 GMT"
},
{
"version": "v4",
"created": "Thu, 10 Apr 2025 14:39:59 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Parihar",
"Rishubh",
""
],
[
"Bhat",
"Abhijnya",
""
],
[
"Basu",
"Abhipsa",
""
],
[
"Mallick",
"Saswat",
""
],
[
"Kundu",
"Jogendra Nath",
""
],
[
"Babu",
"R. Venkatesh",
""
]
] | TITLE: Balancing Act: Distribution-Guided Debiasing in Diffusion Models
ABSTRACT: Diffusion Models (DMs) have emerged as powerful generative models with
unprecedented image generation capability. These models are widely used for
data augmentation and creative applications. However, DMs reflect the biases
present in the training datasets. This is especially concerning in the context
of faces, where the DM prefers one demographic subgroup vs others (eg. female
vs male). In this work, we present a method for debiasing DMs without relying
on additional data or model retraining. Specifically, we propose Distribution
Guidance, which enforces the generated images to follow the prescribed
attribute distribution. To realize this, we build on the key insight that the
latent features of denoising UNet hold rich demographic semantics, and the same
can be leveraged to guide debiased generation. We train Attribute Distribution
Predictor (ADP) - a small mlp that maps the latent features to the distribution
of attributes. ADP is trained with pseudo labels generated from existing
attribute classifiers. The proposed Distribution Guidance with ADP enables us
to do fair generation. Our method reduces bias across single/multiple
attributes and outperforms the baseline by a significant margin for
unconditional and text-conditional diffusion models. Further, we present a
downstream task of training a fair attribute classifier by rebalancing the
training set with our generated data.
|
2402.18307 | Joanne Lin | Joanne Lin, Nantheera Anantrasirichai, David Bull | Multi-Scale Denoising in the Feature Space for Low-Light Instance
Segmentation | Accepted by ICASSP 2025 | 2025 IEEE International Conference on Acoustics, Speech and Signal
Processing (ICASSP), Hyderabad, India, 2025, pp. 1-5 | 10.1109/ICASSP49660.2025.10889336 | ICASSP 2025 | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Instance segmentation for low-light imagery remains largely unexplored due to
the challenges imposed by such conditions, for example shot noise due to low
photon count, color distortions and reduced contrast. In this paper, we propose
an end-to-end solution to address this challenging task. Our proposed method
implements weighted non-local blocks (wNLB) in the feature extractor. This
integration enables an inherent denoising process at the feature level. As a
result, our method eliminates the need for aligned ground truth images during
training, thus supporting training on real-world low-light datasets. We
introduce additional learnable weights at each layer in order to enhance the
network's adaptability to real-world noise characteristics, which affect
different feature scales in different ways. Experimental results on several
object detectors show that the proposed method outperforms the pretrained
networks with an Average Precision (AP) improvement of at least +7.6, with the
introduction of wNLB further enhancing AP by upto +1.3.
| [
{
"version": "v1",
"created": "Wed, 28 Feb 2024 13:07:16 GMT"
},
{
"version": "v2",
"created": "Tue, 10 Sep 2024 08:50:11 GMT"
},
{
"version": "v3",
"created": "Thu, 2 Jan 2025 23:06:37 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Lin",
"Joanne",
""
],
[
"Anantrasirichai",
"Nantheera",
""
],
[
"Bull",
"David",
""
]
] | TITLE: Multi-Scale Denoising in the Feature Space for Low-Light Instance
Segmentation
ABSTRACT: Instance segmentation for low-light imagery remains largely unexplored due to
the challenges imposed by such conditions, for example shot noise due to low
photon count, color distortions and reduced contrast. In this paper, we propose
an end-to-end solution to address this challenging task. Our proposed method
implements weighted non-local blocks (wNLB) in the feature extractor. This
integration enables an inherent denoising process at the feature level. As a
result, our method eliminates the need for aligned ground truth images during
training, thus supporting training on real-world low-light datasets. We
introduce additional learnable weights at each layer in order to enhance the
network's adaptability to real-world noise characteristics, which affect
different feature scales in different ways. Experimental results on several
object detectors show that the proposed method outperforms the pretrained
networks with an Average Precision (AP) improvement of at least +7.6, with the
introduction of wNLB further enhancing AP by upto +1.3.
|
2404.05297 | Sujin Han | Sujin Han, Jinseo Kim, Sung-Ju Lee, Insu Yun | Automated Attack Synthesis for Constant Product Market Makers | 22 pages, 16 figures, 8 tables. Accepted at ACM ISSTA 2025 | null | 10.1145/3728872 | null | cs.CR cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Decentralized Finance (DeFi) enables many novel applications that were
impossible in traditional finances. However, it also introduces new types of
vulnerabilities. An example of such vulnerabilities is a composability bug
between token contracts and Decentralized Exchange (DEX) that follows the
Constant Product Market Maker (CPMM) model. This type of bug, which we refer to
as CPMM composability bug, originates from issues in token contracts that make
them incompatible with CPMMs, thereby endangering other tokens within the CPMM
ecosystem. Since 2022, 23 exploits of such kind have resulted in a total loss
of 2.2M USD. BlockSec, a smart contract auditing company, reported that 138
exploits of such kind occurred just in February 2023.
In this paper, we propose CPMMX , a tool that automatically detects CPMM
composability bugs across entire blockchains. To achieve such scalability, we
first formalized CPMM composability bugs and found that these bugs can be
induced by breaking two safety invariants. Based on this finding, we designed
CPMMX equipped with a two-step approach, called shallow-then-deep search. In
more detail, it first uses shallow search to find transactions that break the
invariants. Then, it uses deep search to refine these transactions, making them
profitable for the attacker. We evaluated CPMMX against five baselines on two
public datasets and one synthetic dataset. In our evaluation, CPMMX detected
2.5x to 1.5x more vulnerabilities compared to baseline methods. It also
analyzed contracts significantly faster, achieving higher F1 scores than the
baselines. Additionally, we applied CPMMX to all contracts on the latest blocks
of the Ethereum and Binance networks and discovered 26 new exploits that can
result in 15.7K USD profit in total.
| [
{
"version": "v1",
"created": "Mon, 8 Apr 2024 08:35:15 GMT"
},
{
"version": "v2",
"created": "Thu, 25 Apr 2024 01:02:53 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 06:19:13 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Han",
"Sujin",
""
],
[
"Kim",
"Jinseo",
""
],
[
"Lee",
"Sung-Ju",
""
],
[
"Yun",
"Insu",
""
]
] | TITLE: Automated Attack Synthesis for Constant Product Market Makers
ABSTRACT: Decentralized Finance (DeFi) enables many novel applications that were
impossible in traditional finances. However, it also introduces new types of
vulnerabilities. An example of such vulnerabilities is a composability bug
between token contracts and Decentralized Exchange (DEX) that follows the
Constant Product Market Maker (CPMM) model. This type of bug, which we refer to
as CPMM composability bug, originates from issues in token contracts that make
them incompatible with CPMMs, thereby endangering other tokens within the CPMM
ecosystem. Since 2022, 23 exploits of such kind have resulted in a total loss
of 2.2M USD. BlockSec, a smart contract auditing company, reported that 138
exploits of such kind occurred just in February 2023.
In this paper, we propose CPMMX , a tool that automatically detects CPMM
composability bugs across entire blockchains. To achieve such scalability, we
first formalized CPMM composability bugs and found that these bugs can be
induced by breaking two safety invariants. Based on this finding, we designed
CPMMX equipped with a two-step approach, called shallow-then-deep search. In
more detail, it first uses shallow search to find transactions that break the
invariants. Then, it uses deep search to refine these transactions, making them
profitable for the attacker. We evaluated CPMMX against five baselines on two
public datasets and one synthetic dataset. In our evaluation, CPMMX detected
2.5x to 1.5x more vulnerabilities compared to baseline methods. It also
analyzed contracts significantly faster, achieving higher F1 scores than the
baselines. Additionally, we applied CPMMX to all contracts on the latest blocks
of the Ethereum and Binance networks and discovered 26 new exploits that can
result in 15.7K USD profit in total.
|
2404.10702 | Arka Ujjal Dey | Arka Ujjal Dey, Artemis Llabr\'es, Ernest Valveny and Dimosthenis
Karatzas | Retrieval Augmented Verification for Zero-Shot Detection of Multimodal
Disinformation | null | null | null | null | cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rise of disinformation on social media, especially through the strategic
manipulation or repurposing of images, paired with provocative text, presents a
complex challenge for traditional fact-checking methods. In this paper, we
introduce a novel zero-shot approach to identify and interpret such multimodal
disinformation, leveraging real-time evidence from credible sources. Our
framework goes beyond simple true-or-false classifications by analyzing both
the textual and visual components of social media claims in a structured,
interpretable manner. By constructing a graph-based representation of entities
and relationships within the claim, combined with pretrained visual features,
our system automatically retrieves and matches external evidence to identify
inconsistencies. Unlike traditional models dependent on labeled datasets, our
method empowers users with transparency, illuminating exactly which aspects of
the claim hold up to scrutiny and which do not. Our framework achieves
competitive performance with state-of-the-art methods while offering enhanced
explainability.
| [
{
"version": "v1",
"created": "Tue, 16 Apr 2024 16:19:22 GMT"
},
{
"version": "v2",
"created": "Mon, 29 Apr 2024 17:19:53 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Apr 2025 22:23:08 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Dey",
"Arka Ujjal",
""
],
[
"Llabrés",
"Artemis",
""
],
[
"Valveny",
"Ernest",
""
],
[
"Karatzas",
"Dimosthenis",
""
]
] | TITLE: Retrieval Augmented Verification for Zero-Shot Detection of Multimodal
Disinformation
ABSTRACT: The rise of disinformation on social media, especially through the strategic
manipulation or repurposing of images, paired with provocative text, presents a
complex challenge for traditional fact-checking methods. In this paper, we
introduce a novel zero-shot approach to identify and interpret such multimodal
disinformation, leveraging real-time evidence from credible sources. Our
framework goes beyond simple true-or-false classifications by analyzing both
the textual and visual components of social media claims in a structured,
interpretable manner. By constructing a graph-based representation of entities
and relationships within the claim, combined with pretrained visual features,
our system automatically retrieves and matches external evidence to identify
inconsistencies. Unlike traditional models dependent on labeled datasets, our
method empowers users with transparency, illuminating exactly which aspects of
the claim hold up to scrutiny and which do not. Our framework achieves
competitive performance with state-of-the-art methods while offering enhanced
explainability.
|
2405.16924 | Francesco Montagna | Francesco Montagna, Max Cairney-Leeming, Dhanya Sridhar, Francesco
Locatello | Demystifying amortized causal discovery with transformers | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Supervised learning approaches for causal discovery from observational data
often achieve competitive performance despite seemingly avoiding explicit
assumptions that traditional methods make for identifiability. In this work, we
investigate CSIvA (Ke et al., 2023), a transformer-based model promising to
train on synthetic data and transfer to real data. First, we bridge the gap
with existing identifiability theory and show that constraints on the training
data distribution implicitly define a prior on the test observations.
Consistent with classical approaches, good performance is achieved when we have
a good prior on the test data, and the underlying model is identifiable. At the
same time, we find new trade-offs. Training on datasets generated from
different classes of causal models, unambiguously identifiable in isolation,
improves the test generalization. Performance is still guaranteed, as the
ambiguous cases resulting from the mixture of identifiable causal models are
unlikely to occur (which we formally prove). Overall, our study finds that
amortized causal discovery still needs to obey identifiability theory, but it
also differs from classical methods in how the assumptions are formulated,
trading more reliance on assumptions on the noise type for fewer hypotheses on
the mechanisms.
| [
{
"version": "v1",
"created": "Mon, 27 May 2024 08:17:49 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Apr 2025 20:30:46 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Montagna",
"Francesco",
""
],
[
"Cairney-Leeming",
"Max",
""
],
[
"Sridhar",
"Dhanya",
""
],
[
"Locatello",
"Francesco",
""
]
] | TITLE: Demystifying amortized causal discovery with transformers
ABSTRACT: Supervised learning approaches for causal discovery from observational data
often achieve competitive performance despite seemingly avoiding explicit
assumptions that traditional methods make for identifiability. In this work, we
investigate CSIvA (Ke et al., 2023), a transformer-based model promising to
train on synthetic data and transfer to real data. First, we bridge the gap
with existing identifiability theory and show that constraints on the training
data distribution implicitly define a prior on the test observations.
Consistent with classical approaches, good performance is achieved when we have
a good prior on the test data, and the underlying model is identifiable. At the
same time, we find new trade-offs. Training on datasets generated from
different classes of causal models, unambiguously identifiable in isolation,
improves the test generalization. Performance is still guaranteed, as the
ambiguous cases resulting from the mixture of identifiable causal models are
unlikely to occur (which we formally prove). Overall, our study finds that
amortized causal discovery still needs to obey identifiability theory, but it
also differs from classical methods in how the assumptions are formulated,
trading more reliance on assumptions on the noise type for fewer hypotheses on
the mechanisms.
|
2405.18560 | Shubhang Bhatnagar | Shubhang Bhatnagar, Narendra Ahuja | Potential Field Based Deep Metric Learning | Accepted to CVPR 2025 | null | null | null | cs.CV cs.AI cs.IR cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep metric learning (DML) involves training a network to learn a
semantically meaningful representation space. Many current approaches mine
n-tuples of examples and model interactions within each tuplets. We present a
novel, compositional DML model that instead of in tuples, represents the
influence of each example (embedding) by a continuous potential field, and
superposes the fields to obtain their combined global potential field. We use
attractive/repulsive potential fields to represent interactions among
embeddings from images of the same/different classes. Contrary to typical
learning methods, where mutual influence of samples is proportional to their
distance, we enforce reduction in such influence with distance, leading to a
decaying field. We show that such decay helps improve performance on real world
datasets with large intra-class variations and label noise. Like other
proxy-based methods, we also use proxies to succinctly represent
sub-populations of examples. We evaluate our method on three standard DML
benchmarks- Cars-196, CUB-200-2011, and SOP datasets where it outperforms
state-of-the-art baselines.
| [
{
"version": "v1",
"created": "Tue, 28 May 2024 20:10:06 GMT"
},
{
"version": "v2",
"created": "Sun, 1 Dec 2024 05:22:22 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 04:49:39 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Bhatnagar",
"Shubhang",
""
],
[
"Ahuja",
"Narendra",
""
]
] | TITLE: Potential Field Based Deep Metric Learning
ABSTRACT: Deep metric learning (DML) involves training a network to learn a
semantically meaningful representation space. Many current approaches mine
n-tuples of examples and model interactions within each tuplets. We present a
novel, compositional DML model that instead of in tuples, represents the
influence of each example (embedding) by a continuous potential field, and
superposes the fields to obtain their combined global potential field. We use
attractive/repulsive potential fields to represent interactions among
embeddings from images of the same/different classes. Contrary to typical
learning methods, where mutual influence of samples is proportional to their
distance, we enforce reduction in such influence with distance, leading to a
decaying field. We show that such decay helps improve performance on real world
datasets with large intra-class variations and label noise. Like other
proxy-based methods, we also use proxies to succinctly represent
sub-populations of examples. We evaluate our method on three standard DML
benchmarks- Cars-196, CUB-200-2011, and SOP datasets where it outperforms
state-of-the-art baselines.
|
2406.07409 | Juntao You | HanQin Cai, Longxiu Huang, Xiliang Lu, Juntao You | Accelerating Ill-conditioned Hankel Matrix Recovery via Structured
Newton-like Descent | null | null | null | null | stat.ML cs.IT cs.LG eess.SP math.IT math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper studies the robust Hankel recovery problem, which simultaneously
removes the sparse outliers and fulfills missing entries from the partial
observation. We propose a novel non-convex algorithm, coined Hankel Structured
Newton-Like Descent (HSNLD), to tackle the robust Hankel recovery problem.
HSNLD is highly efficient with linear convergence, and its convergence rate is
independent of the condition number of the underlying Hankel matrix. The
recovery guarantee has been established under some mild conditions. Numerical
experiments on both synthetic and real datasets show the superior performance
of HSNLD against state-of-the-art algorithms.
| [
{
"version": "v1",
"created": "Tue, 11 Jun 2024 16:14:30 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 07:55:52 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Cai",
"HanQin",
""
],
[
"Huang",
"Longxiu",
""
],
[
"Lu",
"Xiliang",
""
],
[
"You",
"Juntao",
""
]
] | TITLE: Accelerating Ill-conditioned Hankel Matrix Recovery via Structured
Newton-like Descent
ABSTRACT: This paper studies the robust Hankel recovery problem, which simultaneously
removes the sparse outliers and fulfills missing entries from the partial
observation. We propose a novel non-convex algorithm, coined Hankel Structured
Newton-Like Descent (HSNLD), to tackle the robust Hankel recovery problem.
HSNLD is highly efficient with linear convergence, and its convergence rate is
independent of the condition number of the underlying Hankel matrix. The
recovery guarantee has been established under some mild conditions. Numerical
experiments on both synthetic and real datasets show the superior performance
of HSNLD against state-of-the-art algorithms.
|
2406.12123 | Jingxi Xu | Jingxi Xu, Runsheng Wang, Siqi Shang, Ava Chen, Lauren Winterbottom,
To-Liang Hsu, Wenxi Chen, Khondoker Ahmed, Pedro Leandro La Rotta, Xinyue
Zhu, Dawn M. Nilsen, Joel Stein, Matei Ciocarlie | ChatEMG: Synthetic Data Generation to Control a Robotic Hand Orthosis
for Stroke | 8 pages; accepted to RA-L in November 2024; published at RA-L in
February 2025 | null | null | null | cs.RO cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Intent inferral on a hand orthosis for stroke patients is challenging due to
the difficulty of data collection. Additionally, EMG signals exhibit
significant variations across different conditions, sessions, and subjects,
making it hard for classifiers to generalize. Traditional approaches require a
large labeled dataset from the new condition, session, or subject to train
intent classifiers; however, this data collection process is burdensome and
time-consuming. In this paper, we propose ChatEMG, an autoregressive generative
model that can generate synthetic EMG signals conditioned on prompts (i.e., a
given sequence of EMG signals). ChatEMG enables us to collect only a small
dataset from the new condition, session, or subject and expand it with
synthetic samples conditioned on prompts from this new context. ChatEMG
leverages a vast repository of previous data via generative training while
still remaining context-specific via prompting. Our experiments show that these
synthetic samples are classifier-agnostic and can improve intent inferral
accuracy for different types of classifiers. We demonstrate that our complete
approach can be integrated into a single patient session, including the use of
the classifier for functional orthosis-assisted tasks. To the best of our
knowledge, this is the first time an intent classifier trained partially on
synthetic data has been deployed for functional control of an orthosis by a
stroke survivor. Videos, source code, and additional information can be found
at https://jxu.ai/chatemg.
| [
{
"version": "v1",
"created": "Mon, 17 Jun 2024 22:04:44 GMT"
},
{
"version": "v2",
"created": "Fri, 22 Nov 2024 23:15:26 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Apr 2025 21:49:04 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Xu",
"Jingxi",
""
],
[
"Wang",
"Runsheng",
""
],
[
"Shang",
"Siqi",
""
],
[
"Chen",
"Ava",
""
],
[
"Winterbottom",
"Lauren",
""
],
[
"Hsu",
"To-Liang",
""
],
[
"Chen",
"Wenxi",
""
],
[
"Ahmed",
"Khondoker",
""
],
[
"La Rotta",
"Pedro Leandro",
""
],
[
"Zhu",
"Xinyue",
""
],
[
"Nilsen",
"Dawn M.",
""
],
[
"Stein",
"Joel",
""
],
[
"Ciocarlie",
"Matei",
""
]
] | TITLE: ChatEMG: Synthetic Data Generation to Control a Robotic Hand Orthosis
for Stroke
ABSTRACT: Intent inferral on a hand orthosis for stroke patients is challenging due to
the difficulty of data collection. Additionally, EMG signals exhibit
significant variations across different conditions, sessions, and subjects,
making it hard for classifiers to generalize. Traditional approaches require a
large labeled dataset from the new condition, session, or subject to train
intent classifiers; however, this data collection process is burdensome and
time-consuming. In this paper, we propose ChatEMG, an autoregressive generative
model that can generate synthetic EMG signals conditioned on prompts (i.e., a
given sequence of EMG signals). ChatEMG enables us to collect only a small
dataset from the new condition, session, or subject and expand it with
synthetic samples conditioned on prompts from this new context. ChatEMG
leverages a vast repository of previous data via generative training while
still remaining context-specific via prompting. Our experiments show that these
synthetic samples are classifier-agnostic and can improve intent inferral
accuracy for different types of classifiers. We demonstrate that our complete
approach can be integrated into a single patient session, including the use of
the classifier for functional orthosis-assisted tasks. To the best of our
knowledge, this is the first time an intent classifier trained partially on
synthetic data has been deployed for functional control of an orthosis by a
stroke survivor. Videos, source code, and additional information can be found
at https://jxu.ai/chatemg.
|
2406.19388 | Moayed Haji-Ali | Moayed Haji-Ali, Willi Menapace, Aliaksandr Siarohin, Guha
Balakrishnan, Sergey Tulyakov, Vicente Ordonez | Taming Data and Transformers for Scalable Audio Generation | Project Webpage: https://snap-research.github.io/GenAU/ | null | null | null | cs.SD cs.CL cs.CV cs.MM eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The scalability of ambient sound generators is hindered by data scarcity,
insufficient caption quality, and limited scalability in model architecture.
This work addresses these challenges by advancing both data and model scaling.
First, we propose an efficient and scalable dataset collection pipeline
tailored for ambient audio generation, resulting in AutoReCap-XL, the largest
ambient audio-text dataset with over 47 million clips. To provide high-quality
textual annotations, we propose AutoCap, a high-quality automatic audio
captioning model. By adopting a Q-Former module and leveraging audio metadata,
AutoCap substantially enhances caption quality, reaching a CIDEr score of
$83.2$, a $3.2\%$ improvement over previous captioning models. Finally, we
propose GenAu, a scalable transformer-based audio generation architecture that
we scale up to 1.25B parameters. We demonstrate its benefits from data scaling
with synthetic captions as well as model size scaling. When compared to
baseline audio generators trained at similar size and data scale, GenAu obtains
significant improvements of $4.7\%$ in FAD score, $11.1\%$ in IS, and $13.5\%$
in CLAP score. Our code, model checkpoints, and dataset are publicly available.
| [
{
"version": "v1",
"created": "Thu, 27 Jun 2024 17:58:54 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Oct 2024 17:56:21 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 17:55:02 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Haji-Ali",
"Moayed",
""
],
[
"Menapace",
"Willi",
""
],
[
"Siarohin",
"Aliaksandr",
""
],
[
"Balakrishnan",
"Guha",
""
],
[
"Tulyakov",
"Sergey",
""
],
[
"Ordonez",
"Vicente",
""
]
] | TITLE: Taming Data and Transformers for Scalable Audio Generation
ABSTRACT: The scalability of ambient sound generators is hindered by data scarcity,
insufficient caption quality, and limited scalability in model architecture.
This work addresses these challenges by advancing both data and model scaling.
First, we propose an efficient and scalable dataset collection pipeline
tailored for ambient audio generation, resulting in AutoReCap-XL, the largest
ambient audio-text dataset with over 47 million clips. To provide high-quality
textual annotations, we propose AutoCap, a high-quality automatic audio
captioning model. By adopting a Q-Former module and leveraging audio metadata,
AutoCap substantially enhances caption quality, reaching a CIDEr score of
$83.2$, a $3.2\%$ improvement over previous captioning models. Finally, we
propose GenAu, a scalable transformer-based audio generation architecture that
we scale up to 1.25B parameters. We demonstrate its benefits from data scaling
with synthetic captions as well as model size scaling. When compared to
baseline audio generators trained at similar size and data scale, GenAu obtains
significant improvements of $4.7\%$ in FAD score, $11.1\%$ in IS, and $13.5\%$
in CLAP score. Our code, model checkpoints, and dataset are publicly available.
|
2407.15817 | Baudouin Denis de Senneville PhD | Florian Robert, Alexia Calovoulos, Laurent Facq, Fanny Decoeur,
Etienne Gontier, Christophe F. Grosset, Baudouin Denis de Senneville | Enhancing Cell Instance Segmentation in Scanning Electron Microscopy
Images via a Deep Contour Closing Operator | 13 pages, 8 figures, 2 tables | null | 10.1016/j.compbiomed.2025.109972 | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurately segmenting and individualizing cells in SEM images is a highly
promising technique for elucidating tissue architecture in oncology. While
current AI-based methods are effective, errors persist, necessitating
time-consuming manual corrections, particularly in areas where the quality of
cell contours in the image is poor and requires gap filling. This study
presents a novel AI-driven approach for refining cell boundary delineation to
improve instance-based cell segmentation in SEM images, also reducing the
necessity for residual manual correction. A CNN COp-Net is introduced to
address gaps in cell contours, effectively filling in regions with deficient or
absent information. The network takes as input cell contour probability maps
with potentially inadequate or missing information and outputs corrected cell
contour delineations. The lack of training data was addressed by generating low
integrity probability maps using a tailored PDE. We showcase the efficacy of
our approach in augmenting cell boundary precision using both private SEM
images from PDX hepatoblastoma tissues and publicly accessible images datasets.
The proposed cell contour closing operator exhibits a notable improvement in
tested datasets, achieving respectively close to 50% (private data) and 10%
(public data) increase in the accurately-delineated cell proportion compared to
state-of-the-art methods. Additionally, the need for manual corrections was
significantly reduced, therefore facilitating the overall digitalization
process. Our results demonstrate a notable enhancement in the accuracy of cell
instance segmentation, particularly in highly challenging regions where image
quality compromises the integrity of cell boundaries, necessitating gap
filling. Therefore, our work should ultimately facilitate the study of tumour
tissue bioarchitecture in onconanotomy field.
| [
{
"version": "v1",
"created": "Mon, 22 Jul 2024 17:32:06 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 09:20:30 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Robert",
"Florian",
""
],
[
"Calovoulos",
"Alexia",
""
],
[
"Facq",
"Laurent",
""
],
[
"Decoeur",
"Fanny",
""
],
[
"Gontier",
"Etienne",
""
],
[
"Grosset",
"Christophe F.",
""
],
[
"de Senneville",
"Baudouin Denis",
""
]
] | TITLE: Enhancing Cell Instance Segmentation in Scanning Electron Microscopy
Images via a Deep Contour Closing Operator
ABSTRACT: Accurately segmenting and individualizing cells in SEM images is a highly
promising technique for elucidating tissue architecture in oncology. While
current AI-based methods are effective, errors persist, necessitating
time-consuming manual corrections, particularly in areas where the quality of
cell contours in the image is poor and requires gap filling. This study
presents a novel AI-driven approach for refining cell boundary delineation to
improve instance-based cell segmentation in SEM images, also reducing the
necessity for residual manual correction. A CNN COp-Net is introduced to
address gaps in cell contours, effectively filling in regions with deficient or
absent information. The network takes as input cell contour probability maps
with potentially inadequate or missing information and outputs corrected cell
contour delineations. The lack of training data was addressed by generating low
integrity probability maps using a tailored PDE. We showcase the efficacy of
our approach in augmenting cell boundary precision using both private SEM
images from PDX hepatoblastoma tissues and publicly accessible images datasets.
The proposed cell contour closing operator exhibits a notable improvement in
tested datasets, achieving respectively close to 50% (private data) and 10%
(public data) increase in the accurately-delineated cell proportion compared to
state-of-the-art methods. Additionally, the need for manual corrections was
significantly reduced, therefore facilitating the overall digitalization
process. Our results demonstrate a notable enhancement in the accuracy of cell
instance segmentation, particularly in highly challenging regions where image
quality compromises the integrity of cell boundaries, necessitating gap
filling. Therefore, our work should ultimately facilitate the study of tumour
tissue bioarchitecture in onconanotomy field.
|
2409.10365 | Melanie Roschewitz | M\'elanie Roschewitz, Fabio De Sousa Ribeiro, Tian Xia, Galvin Khara,
Ben Glocker | Robust image representations with counterfactual contrastive learning | Code available at
https://github.com/biomedia-mira/counterfactual-contrastive/ | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Contrastive pretraining can substantially increase model generalisation and
downstream performance. However, the quality of the learned representations is
highly dependent on the data augmentation strategy applied to generate positive
pairs. Positive contrastive pairs should preserve semantic meaning while
discarding unwanted variations related to the data acquisition domain.
Traditional contrastive pipelines attempt to simulate domain shifts through
pre-defined generic image transformations. However, these do not always mimic
realistic and relevant domain variations for medical imaging, such as scanner
differences. To tackle this issue, we herein introduce counterfactual
contrastive learning, a novel framework leveraging recent advances in causal
image synthesis to create contrastive positive pairs that faithfully capture
relevant domain variations. Our method, evaluated across five datasets
encompassing both chest radiography and mammography data, for two established
contrastive objectives (SimCLR and DINO-v2), outperforms standard contrastive
learning in terms of robustness to acquisition shift. Notably, counterfactual
contrastive learning achieves superior downstream performance on both
in-distribution and external datasets, especially for images acquired with
scanners under-represented in the training set. Further experiments show that
the proposed framework extends beyond acquisition shifts, with models trained
with counterfactual contrastive learning reducing subgroup disparities across
biological sex.
| [
{
"version": "v1",
"created": "Mon, 16 Sep 2024 15:11:00 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 16:19:20 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Roschewitz",
"Mélanie",
""
],
[
"Ribeiro",
"Fabio De Sousa",
""
],
[
"Xia",
"Tian",
""
],
[
"Khara",
"Galvin",
""
],
[
"Glocker",
"Ben",
""
]
] | TITLE: Robust image representations with counterfactual contrastive learning
ABSTRACT: Contrastive pretraining can substantially increase model generalisation and
downstream performance. However, the quality of the learned representations is
highly dependent on the data augmentation strategy applied to generate positive
pairs. Positive contrastive pairs should preserve semantic meaning while
discarding unwanted variations related to the data acquisition domain.
Traditional contrastive pipelines attempt to simulate domain shifts through
pre-defined generic image transformations. However, these do not always mimic
realistic and relevant domain variations for medical imaging, such as scanner
differences. To tackle this issue, we herein introduce counterfactual
contrastive learning, a novel framework leveraging recent advances in causal
image synthesis to create contrastive positive pairs that faithfully capture
relevant domain variations. Our method, evaluated across five datasets
encompassing both chest radiography and mammography data, for two established
contrastive objectives (SimCLR and DINO-v2), outperforms standard contrastive
learning in terms of robustness to acquisition shift. Notably, counterfactual
contrastive learning achieves superior downstream performance on both
in-distribution and external datasets, especially for images acquired with
scanners under-represented in the training set. Further experiments show that
the proposed framework extends beyond acquisition shifts, with models trained
with counterfactual contrastive learning reducing subgroup disparities across
biological sex.
|
2409.19075 | Jie He | Yu Fu, Jie He, Yifan Yang, Qun Liu, Deyi Xiong | Meta-RTL: Reinforcement-Based Meta-Transfer Learning for Low-Resource
Commonsense Reasoning | There is a text error in table 6 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Meta learning has been widely used to exploit rich-resource source tasks to
improve the performance of low-resource target tasks. Unfortunately, most
existing meta learning approaches treat different source tasks equally,
ignoring the relatedness of source tasks to the target task in knowledge
transfer. To mitigate this issue, we propose a reinforcement-based multi-source
meta-transfer learning framework (Meta-RTL) for low-resource commonsense
reasoning. In this framework, we present a reinforcement-based approach to
dynamically estimating source task weights that measure the contribution of the
corresponding tasks to the target task in the meta-transfer learning. The
differences between the general loss of the meta model and task-specific losses
of source-specific temporal meta models on sampled target data are fed into the
policy network of the reinforcement learning module as rewards. The policy
network is built upon LSTMs that capture long-term dependencies on source task
weight estimation across meta learning iterations. We evaluate the proposed
Meta-RTL using both BERT and ALBERT as the backbone of the meta model on three
commonsense reasoning benchmark datasets. Experimental results demonstrate that
Meta-RTL substantially outperforms strong baselines and previous task selection
strategies and achieves larger improvements on extremely low-resource settings.
| [
{
"version": "v1",
"created": "Fri, 27 Sep 2024 18:22:22 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Mar 2025 09:31:15 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Apr 2025 21:49:23 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Fu",
"Yu",
""
],
[
"He",
"Jie",
""
],
[
"Yang",
"Yifan",
""
],
[
"Liu",
"Qun",
""
],
[
"Xiong",
"Deyi",
""
]
] | TITLE: Meta-RTL: Reinforcement-Based Meta-Transfer Learning for Low-Resource
Commonsense Reasoning
ABSTRACT: Meta learning has been widely used to exploit rich-resource source tasks to
improve the performance of low-resource target tasks. Unfortunately, most
existing meta learning approaches treat different source tasks equally,
ignoring the relatedness of source tasks to the target task in knowledge
transfer. To mitigate this issue, we propose a reinforcement-based multi-source
meta-transfer learning framework (Meta-RTL) for low-resource commonsense
reasoning. In this framework, we present a reinforcement-based approach to
dynamically estimating source task weights that measure the contribution of the
corresponding tasks to the target task in the meta-transfer learning. The
differences between the general loss of the meta model and task-specific losses
of source-specific temporal meta models on sampled target data are fed into the
policy network of the reinforcement learning module as rewards. The policy
network is built upon LSTMs that capture long-term dependencies on source task
weight estimation across meta learning iterations. We evaluate the proposed
Meta-RTL using both BERT and ALBERT as the backbone of the meta model on three
commonsense reasoning benchmark datasets. Experimental results demonstrate that
Meta-RTL substantially outperforms strong baselines and previous task selection
strategies and achieves larger improvements on extremely low-resource settings.
|
2410.01595 | Amin Karimi Monsefi | Pouyan Navard, Amin Karimi Monsefi, Mengxi Zhou, Wei-Lun Chao, Alper
Yilmaz, Rajiv Ramnath | KnobGen: Controlling the Sophistication of Artwork in Sketch-Based
Diffusion Models | Accepted to CVPR 2025 Workshop on CVEU | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recent advances in diffusion models have significantly improved text-to-image
(T2I) generation, but they often struggle to balance fine-grained precision
with high-level control. Methods like ControlNet and T2I-Adapter excel at
following sketches by seasoned artists but tend to be overly rigid, replicating
unintentional flaws in sketches from novice users. Meanwhile, coarse-grained
methods, such as sketch-based abstraction frameworks, offer more accessible
input handling but lack the precise control needed for detailed, professional
use. To address these limitations, we propose KnobGen, a dual-pathway framework
that democratizes sketch-based image generation by seamlessly adapting to
varying levels of sketch complexity and user skill. KnobGen uses a
Coarse-Grained Controller (CGC) module for high-level semantics and a
Fine-Grained Controller (FGC) module for detailed refinement. The relative
strength of these two modules can be adjusted through our knob inference
mechanism to align with the user's specific needs. These mechanisms ensure that
KnobGen can flexibly generate images from both novice sketches and those drawn
by seasoned artists. This maintains control over the final output while
preserving the natural appearance of the image, as evidenced on the
MultiGen-20M dataset and a newly collected sketch dataset.
| [
{
"version": "v1",
"created": "Wed, 2 Oct 2024 14:33:12 GMT"
},
{
"version": "v2",
"created": "Fri, 11 Oct 2024 12:47:48 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Apr 2025 22:27:10 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Navard",
"Pouyan",
""
],
[
"Monsefi",
"Amin Karimi",
""
],
[
"Zhou",
"Mengxi",
""
],
[
"Chao",
"Wei-Lun",
""
],
[
"Yilmaz",
"Alper",
""
],
[
"Ramnath",
"Rajiv",
""
]
] | TITLE: KnobGen: Controlling the Sophistication of Artwork in Sketch-Based
Diffusion Models
ABSTRACT: Recent advances in diffusion models have significantly improved text-to-image
(T2I) generation, but they often struggle to balance fine-grained precision
with high-level control. Methods like ControlNet and T2I-Adapter excel at
following sketches by seasoned artists but tend to be overly rigid, replicating
unintentional flaws in sketches from novice users. Meanwhile, coarse-grained
methods, such as sketch-based abstraction frameworks, offer more accessible
input handling but lack the precise control needed for detailed, professional
use. To address these limitations, we propose KnobGen, a dual-pathway framework
that democratizes sketch-based image generation by seamlessly adapting to
varying levels of sketch complexity and user skill. KnobGen uses a
Coarse-Grained Controller (CGC) module for high-level semantics and a
Fine-Grained Controller (FGC) module for detailed refinement. The relative
strength of these two modules can be adjusted through our knob inference
mechanism to align with the user's specific needs. These mechanisms ensure that
KnobGen can flexibly generate images from both novice sketches and those drawn
by seasoned artists. This maintains control over the final output while
preserving the natural appearance of the image, as evidenced on the
MultiGen-20M dataset and a newly collected sketch dataset.
|
2410.03749 | Larry Liebovitch | K. Lian (1), L. S. Liebovitch (1), M. Wild (1), H. West (1), P. T.
Coleman (1), F. Chen (2), E. Kimani (2), K. Sieck (2) ((1) Columbia
University, (2) Toyota Research Institute) | Machine Learning Classification of Peaceful Countries: A Comparative
Analysis and Dataset Optimization | 5 pages, 5 figures | 2025 59th Annual Conference on Information Sciences and Systems
(CISS), Baltimore, MD, USA, 2025, pp. 1-5 | 10.1109/CISS64860.2025.10944706 | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper presents a machine learning approach to classify countries as
peaceful or non-peaceful using linguistic patterns extracted from global media
articles. We employ vector embeddings and cosine similarity to develop a
supervised classification model that effectively identifies peaceful countries.
Additionally, we explore the impact of dataset size on model performance,
investigating how shrinking the dataset influences classification accuracy. Our
results highlight the challenges and opportunities associated with using
large-scale text data for peace studies.
| [
{
"version": "v1",
"created": "Tue, 1 Oct 2024 19:28:03 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Lian",
"K.",
""
],
[
"Liebovitch",
"L. S.",
""
],
[
"Wild",
"M.",
""
],
[
"West",
"H.",
""
],
[
"Coleman",
"P. T.",
""
],
[
"Chen",
"F.",
""
],
[
"Kimani",
"E.",
""
],
[
"Sieck",
"K.",
""
]
] | TITLE: Machine Learning Classification of Peaceful Countries: A Comparative
Analysis and Dataset Optimization
ABSTRACT: This paper presents a machine learning approach to classify countries as
peaceful or non-peaceful using linguistic patterns extracted from global media
articles. We employ vector embeddings and cosine similarity to develop a
supervised classification model that effectively identifies peaceful countries.
Additionally, we explore the impact of dataset size on model performance,
investigating how shrinking the dataset influences classification accuracy. Our
results highlight the challenges and opportunities associated with using
large-scale text data for peace studies.
|
2410.08206 | Idil Esen Zulfikar | Ilya Fradlin, Idil Esen Zulfikar, Kadir Yilmaz, Theodora Kontogianni,
Bastian Leibe | Interactive4D: Interactive 4D LiDAR Segmentation | Accepted to ICRA2025! | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Interactive segmentation has an important role in facilitating the annotation
process of future LiDAR datasets. Existing approaches sequentially segment
individual objects at each LiDAR scan, repeating the process throughout the
entire sequence, which is redundant and ineffective. In this work, we propose
interactive 4D segmentation, a new paradigm that allows segmenting multiple
objects on multiple LiDAR scans simultaneously, and Interactive4D, the first
interactive 4D segmentation model that segments multiple objects on
superimposed consecutive LiDAR scans in a single iteration by utilizing the
sequential nature of LiDAR data. While performing interactive segmentation, our
model leverages the entire space-time volume, leading to more efficient
segmentation. Operating on the 4D volume, it directly provides consistent
instance IDs over time and also simplifies tracking annotations. Moreover, we
show that click simulations are crucial for successful model training on LiDAR
point clouds. To this end, we design a click simulation strategy that is better
suited for the characteristics of LiDAR data. To demonstrate its accuracy and
effectiveness, we evaluate Interactive4D on multiple LiDAR datasets, where
Interactive4D achieves a new state-of-the-art by a large margin. We publicly
release the code and models at https://vision.rwth-aachen.de/Interactive4D.
| [
{
"version": "v1",
"created": "Thu, 10 Oct 2024 17:59:45 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 17:59:53 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Fradlin",
"Ilya",
""
],
[
"Zulfikar",
"Idil Esen",
""
],
[
"Yilmaz",
"Kadir",
""
],
[
"Kontogianni",
"Theodora",
""
],
[
"Leibe",
"Bastian",
""
]
] | TITLE: Interactive4D: Interactive 4D LiDAR Segmentation
ABSTRACT: Interactive segmentation has an important role in facilitating the annotation
process of future LiDAR datasets. Existing approaches sequentially segment
individual objects at each LiDAR scan, repeating the process throughout the
entire sequence, which is redundant and ineffective. In this work, we propose
interactive 4D segmentation, a new paradigm that allows segmenting multiple
objects on multiple LiDAR scans simultaneously, and Interactive4D, the first
interactive 4D segmentation model that segments multiple objects on
superimposed consecutive LiDAR scans in a single iteration by utilizing the
sequential nature of LiDAR data. While performing interactive segmentation, our
model leverages the entire space-time volume, leading to more efficient
segmentation. Operating on the 4D volume, it directly provides consistent
instance IDs over time and also simplifies tracking annotations. Moreover, we
show that click simulations are crucial for successful model training on LiDAR
point clouds. To this end, we design a click simulation strategy that is better
suited for the characteristics of LiDAR data. To demonstrate its accuracy and
effectiveness, we evaluate Interactive4D on multiple LiDAR datasets, where
Interactive4D achieves a new state-of-the-art by a large margin. We publicly
release the code and models at https://vision.rwth-aachen.de/Interactive4D.
|
2410.13453 | Ant Duru | Ant Duru, Alptekin Temizel | Adaptive Augmentation Policy Optimization with LLM Feedback | 15 pages, 4 tables, 3 figures submitted for consideration to 2025
Medical Image Understanding and Analysis Conference (MIUA) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data augmentation is a critical component of deep learning pipelines,
enhancing model generalization by increasing dataset diversity. Traditional
augmentation strategies rely on manually designed transformations, stochastic
sampling, or automated search-based approaches. Although automated methods
improve performance, they often require extensive computational resources and
are tailored to specific datasets. In this work, we propose a Large Language
Model (LLM)-guided augmentation optimization strategy that refines augmentation
policies based on model performance feedback. We introduce two approaches: (1)
LLM-Guided Augmentation Policy Optimization, where augmentation policies are
selected by an LLM prior to training and iteratively refined across multiple
training cycles, and (2) Adaptive LLM-Guided Augmentation Policy Optimization,
where policies adapt in real-time based on performance metrics. This
in-training approach eliminates the need for full model retraining before
receiving LLM feedback, thereby reducing computational costs while improving
performance. Our methodology employs an LLM to dynamically select augmentation
transformations based on dataset characteristics, model architecture, and prior
training outcomes. Unlike traditional search-based methods, our approach
leverages the contextual knowledge of LLMs, particularly in specialized domains
like medical imaging, to recommend augmentation strategies tailored to
domain-specific data. We evaluate our approach on multiple domain-specific
image classification datasets where augmentation is key to model robustness.
Results show that LLM-guided augmentation optimization outperforms traditional
methods, improving model accuracy. These findings highlight the potential of
LLMs in automating and adapting deep learning training workflows.
| [
{
"version": "v1",
"created": "Thu, 17 Oct 2024 11:26:10 GMT"
},
{
"version": "v2",
"created": "Tue, 8 Apr 2025 11:05:01 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Apr 2025 18:00:00 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Duru",
"Ant",
""
],
[
"Temizel",
"Alptekin",
""
]
] | TITLE: Adaptive Augmentation Policy Optimization with LLM Feedback
ABSTRACT: Data augmentation is a critical component of deep learning pipelines,
enhancing model generalization by increasing dataset diversity. Traditional
augmentation strategies rely on manually designed transformations, stochastic
sampling, or automated search-based approaches. Although automated methods
improve performance, they often require extensive computational resources and
are tailored to specific datasets. In this work, we propose a Large Language
Model (LLM)-guided augmentation optimization strategy that refines augmentation
policies based on model performance feedback. We introduce two approaches: (1)
LLM-Guided Augmentation Policy Optimization, where augmentation policies are
selected by an LLM prior to training and iteratively refined across multiple
training cycles, and (2) Adaptive LLM-Guided Augmentation Policy Optimization,
where policies adapt in real-time based on performance metrics. This
in-training approach eliminates the need for full model retraining before
receiving LLM feedback, thereby reducing computational costs while improving
performance. Our methodology employs an LLM to dynamically select augmentation
transformations based on dataset characteristics, model architecture, and prior
training outcomes. Unlike traditional search-based methods, our approach
leverages the contextual knowledge of LLMs, particularly in specialized domains
like medical imaging, to recommend augmentation strategies tailored to
domain-specific data. We evaluate our approach on multiple domain-specific
image classification datasets where augmentation is key to model robustness.
Results show that LLM-guided augmentation optimization outperforms traditional
methods, improving model accuracy. These findings highlight the potential of
LLMs in automating and adapting deep learning training workflows.
|
2410.22318 | Can Chen | Can Chen, Jun-Kun Wang | Online Detecting LLM-Generated Texts via Sequential Hypothesis Testing
by Betting | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Developing algorithms to differentiate between machine-generated texts and
human-written texts has garnered substantial attention in recent years.
Existing methods in this direction typically concern an offline setting where a
dataset containing a mix of real and machine-generated texts is given upfront,
and the task is to determine whether each sample in the dataset is from a large
language model (LLM) or a human. However, in many practical scenarios, sources
such as news websites, social media accounts, or on other forums publish
content in a streaming fashion. Therefore, in this online scenario, how to
quickly and accurately determine whether the source is an LLM with strong
statistical guarantees is crucial for these media or platforms to function
effectively and prevent the spread of misinformation and other potential misuse
of LLMs. To tackle the problem of online detection, we develop an algorithm
based on the techniques of sequential hypothesis testing by betting that not
only builds upon and complements existing offline detection techniques but also
enjoys statistical guarantees, which include a controlled false positive rate
and the expected time to correctly identify a source as an LLM. Experiments
were conducted to demonstrate the effectiveness of our method.
| [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 17:55:14 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 00:51:20 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Chen",
"Can",
""
],
[
"Wang",
"Jun-Kun",
""
]
] | TITLE: Online Detecting LLM-Generated Texts via Sequential Hypothesis Testing
by Betting
ABSTRACT: Developing algorithms to differentiate between machine-generated texts and
human-written texts has garnered substantial attention in recent years.
Existing methods in this direction typically concern an offline setting where a
dataset containing a mix of real and machine-generated texts is given upfront,
and the task is to determine whether each sample in the dataset is from a large
language model (LLM) or a human. However, in many practical scenarios, sources
such as news websites, social media accounts, or on other forums publish
content in a streaming fashion. Therefore, in this online scenario, how to
quickly and accurately determine whether the source is an LLM with strong
statistical guarantees is crucial for these media or platforms to function
effectively and prevent the spread of misinformation and other potential misuse
of LLMs. To tackle the problem of online detection, we develop an algorithm
based on the techniques of sequential hypothesis testing by betting that not
only builds upon and complements existing offline detection techniques but also
enjoys statistical guarantees, which include a controlled false positive rate
and the expected time to correctly identify a source as an LLM. Experiments
were conducted to demonstrate the effectiveness of our method.
|
2410.23780 | Maixuan Xue | Xinyuan Chang, Maixuan Xue, Xinran Liu, Zheng Pan, Xing Wei | Driving by the Rules: A Benchmark for Integrating Traffic Sign
Regulations into Vectorized HD Map | 26 pages, 16 figures. Accepted as a Highlight at CVPR 2025. Project
page: https://miv-xjtu.github.io/MapDR/ | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Ensuring adherence to traffic sign regulations is essential for both human
and autonomous vehicle navigation. While current online mapping solutions often
prioritize the construction of the geometric and connectivity layers of HD
maps, overlooking the construction of the traffic regulation layer within HD
maps. Addressing this gap, we introduce MapDR, a novel dataset designed for the
extraction of Driving Rules from traffic signs and their association with
vectorized, locally perceived HD Maps. MapDR features over $10,000$ annotated
video clips that capture the intricate correlation between traffic sign
regulations and lanes. Built upon this benchmark and the newly defined task of
integrating traffic regulations into online HD maps, we provide modular and
end-to-end solutions: VLE-MEE and RuleVLM, offering a strong baseline for
advancing autonomous driving technology. It fills a critical gap in the
integration of traffic sign rules, contributing to the development of reliable
autonomous driving systems. Code is available at
https://github.com/MIV-XJTU/MapDR.
| [
{
"version": "v1",
"created": "Thu, 31 Oct 2024 09:53:21 GMT"
},
{
"version": "v2",
"created": "Mon, 6 Jan 2025 12:07:55 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 11:13:00 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Chang",
"Xinyuan",
""
],
[
"Xue",
"Maixuan",
""
],
[
"Liu",
"Xinran",
""
],
[
"Pan",
"Zheng",
""
],
[
"Wei",
"Xing",
""
]
] | TITLE: Driving by the Rules: A Benchmark for Integrating Traffic Sign
Regulations into Vectorized HD Map
ABSTRACT: Ensuring adherence to traffic sign regulations is essential for both human
and autonomous vehicle navigation. While current online mapping solutions often
prioritize the construction of the geometric and connectivity layers of HD
maps, overlooking the construction of the traffic regulation layer within HD
maps. Addressing this gap, we introduce MapDR, a novel dataset designed for the
extraction of Driving Rules from traffic signs and their association with
vectorized, locally perceived HD Maps. MapDR features over $10,000$ annotated
video clips that capture the intricate correlation between traffic sign
regulations and lanes. Built upon this benchmark and the newly defined task of
integrating traffic regulations into online HD maps, we provide modular and
end-to-end solutions: VLE-MEE and RuleVLM, offering a strong baseline for
advancing autonomous driving technology. It fills a critical gap in the
integration of traffic sign rules, contributing to the development of reliable
autonomous driving systems. Code is available at
https://github.com/MIV-XJTU/MapDR.
|
2411.08033 | Yushi Lan | Yushi Lan, Shangchen Zhou, Zhaoyang Lyu, Fangzhou Hong, Shuai Yang, Bo
Dai, Xingang Pan, Chen Change Loy | GaussianAnything: Interactive Point Cloud Flow Matching For 3D Object
Generation | ICLR 2025 project page: https://nirvanalan.github.io/projects/GA/ | null | null | null | cs.CV cs.AI cs.GR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | While 3D content generation has advanced significantly, existing methods
still face challenges with input formats, latent space design, and output
representations. This paper introduces a novel 3D generation framework that
addresses these challenges, offering scalable, high-quality 3D generation with
an interactive Point Cloud-structured Latent space. Our framework employs a
Variational Autoencoder (VAE) with multi-view posed RGB-D(epth)-N(ormal)
renderings as input, using a unique latent space design that preserves 3D shape
information, and incorporates a cascaded latent flow-based model for improved
shape-texture disentanglement. The proposed method, GaussianAnything, supports
multi-modal conditional 3D generation, allowing for point cloud, caption, and
single image inputs. Notably, the newly proposed latent space naturally enables
geometry-texture disentanglement, thus allowing 3D-aware editing. Experimental
results demonstrate the effectiveness of our approach on multiple datasets,
outperforming existing native 3D methods in both text- and image-conditioned 3D
generation.
| [
{
"version": "v1",
"created": "Tue, 12 Nov 2024 18:59:32 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 12:24:52 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Lan",
"Yushi",
""
],
[
"Zhou",
"Shangchen",
""
],
[
"Lyu",
"Zhaoyang",
""
],
[
"Hong",
"Fangzhou",
""
],
[
"Yang",
"Shuai",
""
],
[
"Dai",
"Bo",
""
],
[
"Pan",
"Xingang",
""
],
[
"Loy",
"Chen Change",
""
]
] | TITLE: GaussianAnything: Interactive Point Cloud Flow Matching For 3D Object
Generation
ABSTRACT: While 3D content generation has advanced significantly, existing methods
still face challenges with input formats, latent space design, and output
representations. This paper introduces a novel 3D generation framework that
addresses these challenges, offering scalable, high-quality 3D generation with
an interactive Point Cloud-structured Latent space. Our framework employs a
Variational Autoencoder (VAE) with multi-view posed RGB-D(epth)-N(ormal)
renderings as input, using a unique latent space design that preserves 3D shape
information, and incorporates a cascaded latent flow-based model for improved
shape-texture disentanglement. The proposed method, GaussianAnything, supports
multi-modal conditional 3D generation, allowing for point cloud, caption, and
single image inputs. Notably, the newly proposed latent space naturally enables
geometry-texture disentanglement, thus allowing 3D-aware editing. Experimental
results demonstrate the effectiveness of our approach on multiple datasets,
outperforming existing native 3D methods in both text- and image-conditioned 3D
generation.
|
2411.08753 | Sagnik Majumder | Sagnik Majumder, Tushar Nagarajan, Ziad Al-Halah, Reina Pradhan,
Kristen Grauman | Which Viewpoint Shows it Best? Language for Weakly Supervising View
Selection in Multi-view Instructional Videos | Accepted to CVPR 2025 (Highlight) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a multi-view video, which viewpoint is most informative for a human
observer? Existing methods rely on heuristics or expensive "best-view"
supervision to answer this question, limiting their applicability. We propose a
weakly supervised approach that leverages language accompanying an
instructional multi-view video as a means to recover its most informative
viewpoint(s). Our key hypothesis is that the more accurately an individual view
can predict a view-agnostic text summary, the more informative it is. To put
this into action, we propose LangView, a framework that uses the relative
accuracy of view-dependent caption predictions as a proxy for best view
pseudo-labels. Then, those pseudo-labels are used to train a view selector,
together with an auxiliary camera pose predictor that enhances
view-sensitivity. During inference, our model takes as input only a multi-view
video--no language or camera poses--and returns the best viewpoint to watch at
each timestep. On two challenging datasets comprised of diverse multi-camera
setups and how-to activities, our model consistently outperforms
state-of-the-art baselines, both with quantitative metrics and human
evaluation. Project page:
https://vision.cs.utexas.edu/projects/which-view-shows-it-best.
| [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 16:31:08 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Dec 2024 15:49:20 GMT"
},
{
"version": "v3",
"created": "Fri, 4 Apr 2025 20:45:11 GMT"
},
{
"version": "v4",
"created": "Thu, 10 Apr 2025 02:02:49 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Majumder",
"Sagnik",
""
],
[
"Nagarajan",
"Tushar",
""
],
[
"Al-Halah",
"Ziad",
""
],
[
"Pradhan",
"Reina",
""
],
[
"Grauman",
"Kristen",
""
]
] | TITLE: Which Viewpoint Shows it Best? Language for Weakly Supervising View
Selection in Multi-view Instructional Videos
ABSTRACT: Given a multi-view video, which viewpoint is most informative for a human
observer? Existing methods rely on heuristics or expensive "best-view"
supervision to answer this question, limiting their applicability. We propose a
weakly supervised approach that leverages language accompanying an
instructional multi-view video as a means to recover its most informative
viewpoint(s). Our key hypothesis is that the more accurately an individual view
can predict a view-agnostic text summary, the more informative it is. To put
this into action, we propose LangView, a framework that uses the relative
accuracy of view-dependent caption predictions as a proxy for best view
pseudo-labels. Then, those pseudo-labels are used to train a view selector,
together with an auxiliary camera pose predictor that enhances
view-sensitivity. During inference, our model takes as input only a multi-view
video--no language or camera poses--and returns the best viewpoint to watch at
each timestep. On two challenging datasets comprised of diverse multi-camera
setups and how-to activities, our model consistently outperforms
state-of-the-art baselines, both with quantitative metrics and human
evaluation. Project page:
https://vision.cs.utexas.edu/projects/which-view-shows-it-best.
|
2411.10739 | Jiangang Chen | Jiangang Chen, Yung-Hong Sun, Kristen Pickett, Barbara King, Yu Hen
Hu, Hongrui Jiang | A Wearable Gait Monitoring System for 17 Gait Parameters Based on
Computer Vision | 13 pages, 14 figures. This paper was submitted for publication to the
IEEE Transactions on Instrumentation and Measurement | null | 10.1109/TIM.2025.3557814 | null | eess.SY cs.CV cs.SY eess.SP | http://creativecommons.org/licenses/by/4.0/ | We developed a shoe-mounted gait monitoring system capable of tracking up to
17 gait parameters, including gait length, step time, stride velocity, and
others. The system employs a stereo camera mounted on one shoe to track a
marker placed on the opposite shoe, enabling the estimation of spatial gait
parameters. Additionally, a Force Sensitive Resistor (FSR) affixed to the heel
of the shoe, combined with a custom-designed algorithm, is utilized to measure
temporal gait parameters. Through testing on multiple participants and
comparison with the gait mat, the proposed gait monitoring system exhibited
notable performance, with the accuracy of all measured gait parameters
exceeding 93.61%. The system also demonstrated a low drift of 4.89% during
long-distance walking. A gait identification task conducted on participants
using a trained Transformer model achieved 95.7% accuracy on the dataset
collected by the proposed system, demonstrating that our hardware has the
potential to collect long-sequence gait data suitable for integration with
current Large Language Models (LLMs). The system is cost-effective,
user-friendly, and well-suited for real-life measurements.
| [
{
"version": "v1",
"created": "Sat, 16 Nov 2024 08:25:22 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Chen",
"Jiangang",
""
],
[
"Sun",
"Yung-Hong",
""
],
[
"Pickett",
"Kristen",
""
],
[
"King",
"Barbara",
""
],
[
"Hu",
"Yu Hen",
""
],
[
"Jiang",
"Hongrui",
""
]
] | TITLE: A Wearable Gait Monitoring System for 17 Gait Parameters Based on
Computer Vision
ABSTRACT: We developed a shoe-mounted gait monitoring system capable of tracking up to
17 gait parameters, including gait length, step time, stride velocity, and
others. The system employs a stereo camera mounted on one shoe to track a
marker placed on the opposite shoe, enabling the estimation of spatial gait
parameters. Additionally, a Force Sensitive Resistor (FSR) affixed to the heel
of the shoe, combined with a custom-designed algorithm, is utilized to measure
temporal gait parameters. Through testing on multiple participants and
comparison with the gait mat, the proposed gait monitoring system exhibited
notable performance, with the accuracy of all measured gait parameters
exceeding 93.61%. The system also demonstrated a low drift of 4.89% during
long-distance walking. A gait identification task conducted on participants
using a trained Transformer model achieved 95.7% accuracy on the dataset
collected by the proposed system, demonstrating that our hardware has the
potential to collect long-sequence gait data suitable for integration with
current Large Language Models (LLMs). The system is cost-effective,
user-friendly, and well-suited for real-life measurements.
|
2411.11260 | Cameron Morin | Cameron Morin and Matti Marttinen Larsson | Large corpora and large language models: a replicable method for
automating grammatical annotation | null | Linguistics Vanguard, 1-10 (2025) | 10.1515/lingvan-2024-0228 | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Much linguistic research relies on annotated datasets of features extracted
from text corpora, but the rapid quantitative growth of these corpora has
created practical difficulties for linguists to manually annotate large data
samples. In this paper, we present a replicable, supervised method that
leverages large language models for assisting the linguist in grammatical
annotation through prompt engineering, training, and evaluation. We introduce a
methodological pipeline applied to the case study of formal variation in the
English evaluative verb construction 'consider X (as) (to be) Y', based on the
large language model Claude 3.5 Sonnet and corpus data from Davies' NOW and
EnTenTen21 (SketchEngine). Overall, we reach a model accuracy of over 90% on
our held-out test samples with only a small amount of training data, validating
the method for the annotation of very large quantities of tokens of the
construction in the future. We discuss the generalisability of our results for
a wider range of case studies of grammatical constructions and grammatical
variation and change, underlining the value of AI copilots as tools for future
linguistic research, notwithstanding some important caveats.
| [
{
"version": "v1",
"created": "Mon, 18 Nov 2024 03:29:48 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 07:24:50 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Morin",
"Cameron",
""
],
[
"Larsson",
"Matti Marttinen",
""
]
] | TITLE: Large corpora and large language models: a replicable method for
automating grammatical annotation
ABSTRACT: Much linguistic research relies on annotated datasets of features extracted
from text corpora, but the rapid quantitative growth of these corpora has
created practical difficulties for linguists to manually annotate large data
samples. In this paper, we present a replicable, supervised method that
leverages large language models for assisting the linguist in grammatical
annotation through prompt engineering, training, and evaluation. We introduce a
methodological pipeline applied to the case study of formal variation in the
English evaluative verb construction 'consider X (as) (to be) Y', based on the
large language model Claude 3.5 Sonnet and corpus data from Davies' NOW and
EnTenTen21 (SketchEngine). Overall, we reach a model accuracy of over 90% on
our held-out test samples with only a small amount of training data, validating
the method for the annotation of very large quantities of tokens of the
construction in the future. We discuss the generalisability of our results for
a wider range of case studies of grammatical constructions and grammatical
variation and change, underlining the value of AI copilots as tools for future
linguistic research, notwithstanding some important caveats.
|
2411.15139 | Bencheng Liao | Bencheng Liao, Shaoyu Chen, Haoran Yin, Bo Jiang, Cheng Wang, Sixu
Yan, Xinbang Zhang, Xiangyu Li, Ying Zhang, Qian Zhang, Xinggang Wang | DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous
Driving | Accepted to CVPR 2025 as Highlight. Code & demo & model are available
at https://github.com/hustvl/DiffusionDrive | null | null | null | cs.CV cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recently, the diffusion model has emerged as a powerful generative technique
for robotic policy learning, capable of modeling multi-mode action
distributions. Leveraging its capability for end-to-end autonomous driving is a
promising direction. However, the numerous denoising steps in the robotic
diffusion policy and the more dynamic, open-world nature of traffic scenes pose
substantial challenges for generating diverse driving actions at a real-time
speed. To address these challenges, we propose a novel truncated diffusion
policy that incorporates prior multi-mode anchors and truncates the diffusion
schedule, enabling the model to learn denoising from anchored Gaussian
distribution to the multi-mode driving action distribution. Additionally, we
design an efficient cascade diffusion decoder for enhanced interaction with
conditional scene context. The proposed model, DiffusionDrive, demonstrates
10$\times$ reduction in denoising steps compared to vanilla diffusion policy,
delivering superior diversity and quality in just 2 steps. On the
planning-oriented NAVSIM dataset, with the aligned ResNet-34 backbone,
DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new
record, while running at a real-time speed of 45 FPS on an NVIDIA 4090.
Qualitative results on challenging scenarios further confirm that
DiffusionDrive can robustly generate diverse plausible driving actions. Code
and model will be available at https://github.com/hustvl/DiffusionDrive.
| [
{
"version": "v1",
"created": "Fri, 22 Nov 2024 18:59:47 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Mar 2025 03:02:15 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 08:48:37 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Liao",
"Bencheng",
""
],
[
"Chen",
"Shaoyu",
""
],
[
"Yin",
"Haoran",
""
],
[
"Jiang",
"Bo",
""
],
[
"Wang",
"Cheng",
""
],
[
"Yan",
"Sixu",
""
],
[
"Zhang",
"Xinbang",
""
],
[
"Li",
"Xiangyu",
""
],
[
"Zhang",
"Ying",
""
],
[
"Zhang",
"Qian",
""
],
[
"Wang",
"Xinggang",
""
]
] | TITLE: DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous
Driving
ABSTRACT: Recently, the diffusion model has emerged as a powerful generative technique
for robotic policy learning, capable of modeling multi-mode action
distributions. Leveraging its capability for end-to-end autonomous driving is a
promising direction. However, the numerous denoising steps in the robotic
diffusion policy and the more dynamic, open-world nature of traffic scenes pose
substantial challenges for generating diverse driving actions at a real-time
speed. To address these challenges, we propose a novel truncated diffusion
policy that incorporates prior multi-mode anchors and truncates the diffusion
schedule, enabling the model to learn denoising from anchored Gaussian
distribution to the multi-mode driving action distribution. Additionally, we
design an efficient cascade diffusion decoder for enhanced interaction with
conditional scene context. The proposed model, DiffusionDrive, demonstrates
10$\times$ reduction in denoising steps compared to vanilla diffusion policy,
delivering superior diversity and quality in just 2 steps. On the
planning-oriented NAVSIM dataset, with the aligned ResNet-34 backbone,
DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new
record, while running at a real-time speed of 45 FPS on an NVIDIA 4090.
Qualitative results on challenging scenarios further confirm that
DiffusionDrive can robustly generate diverse plausible driving actions. Code
and model will be available at https://github.com/hustvl/DiffusionDrive.
|
2411.17060 | Mark Iskarous | Mark M. Iskarous, Zan Chaudhry, Fangjie Li, Samuel Bello, Sriramana
Sankar, Ariel Slepyan, Natasha Chugh, Christopher L. Hunt, Rebecca J. Greene,
Nitish V. Thakor | Invariant neuromorphic representations of tactile stimuli improve
robustness of a real-time texture classification system | 34 pages, 9 figures, 1 table | null | 10.1002/aisy.202401078 | null | cs.RO eess.SP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Humans have an exquisite sense of touch which robotic and prosthetic systems
aim to recreate. We developed algorithms to create neuron-like (neuromorphic)
spiking representations of texture that are invariant to the scanning speed and
contact force applied in the sensing process. The spiking representations are
based on mimicking activity from mechanoreceptors in human skin and further
processing up to the brain. The neuromorphic encoding process transforms analog
sensor readings into speed and force invariant spiking representations in three
sequential stages: the force invariance module (in the analog domain), the
spiking activity encoding module (transforms from analog to spiking domain),
and the speed invariance module (in the spiking domain). The algorithms were
tested on a tactile texture dataset collected in 15 speed-force conditions. An
offline texture classification system built on the invariant representations
has higher classification accuracy, improved computational efficiency, and
increased capability to identify textures explored in novel speed-force
conditions. The speed invariance algorithm was adapted to a real-time
human-operated texture classification system. Similarly, the invariant
representations improved classification accuracy, computational efficiency, and
capability to identify textures explored in novel conditions. The invariant
representation is even more crucial in this context due to human imprecision
which seems to the classification system as a novel condition. These results
demonstrate that invariant neuromorphic representations enable better
performing neurorobotic tactile sensing systems. Furthermore, because the
neuromorphic representations are based on biological processing, this work can
be used in the future as the basis for naturalistic sensory feedback for upper
limb amputees.
| [
{
"version": "v1",
"created": "Tue, 26 Nov 2024 02:57:37 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Iskarous",
"Mark M.",
""
],
[
"Chaudhry",
"Zan",
""
],
[
"Li",
"Fangjie",
""
],
[
"Bello",
"Samuel",
""
],
[
"Sankar",
"Sriramana",
""
],
[
"Slepyan",
"Ariel",
""
],
[
"Chugh",
"Natasha",
""
],
[
"Hunt",
"Christopher L.",
""
],
[
"Greene",
"Rebecca J.",
""
],
[
"Thakor",
"Nitish V.",
""
]
] | TITLE: Invariant neuromorphic representations of tactile stimuli improve
robustness of a real-time texture classification system
ABSTRACT: Humans have an exquisite sense of touch which robotic and prosthetic systems
aim to recreate. We developed algorithms to create neuron-like (neuromorphic)
spiking representations of texture that are invariant to the scanning speed and
contact force applied in the sensing process. The spiking representations are
based on mimicking activity from mechanoreceptors in human skin and further
processing up to the brain. The neuromorphic encoding process transforms analog
sensor readings into speed and force invariant spiking representations in three
sequential stages: the force invariance module (in the analog domain), the
spiking activity encoding module (transforms from analog to spiking domain),
and the speed invariance module (in the spiking domain). The algorithms were
tested on a tactile texture dataset collected in 15 speed-force conditions. An
offline texture classification system built on the invariant representations
has higher classification accuracy, improved computational efficiency, and
increased capability to identify textures explored in novel speed-force
conditions. The speed invariance algorithm was adapted to a real-time
human-operated texture classification system. Similarly, the invariant
representations improved classification accuracy, computational efficiency, and
capability to identify textures explored in novel conditions. The invariant
representation is even more crucial in this context due to human imprecision
which seems to the classification system as a novel condition. These results
demonstrate that invariant neuromorphic representations enable better
performing neurorobotic tactile sensing systems. Furthermore, because the
neuromorphic representations are based on biological processing, this work can
be used in the future as the basis for naturalistic sensory feedback for upper
limb amputees.
|
2411.19050 | Nicola Fanelli | Nicola Fanelli, Gennaro Vessio, Giovanna Castellano | I Dream My Painting: Connecting MLLMs and Diffusion Models via Prompt
Generation for Text-Guided Multi-Mask Inpainting | Accepted at WACV 2025 | null | 10.1109/WACV61041.2025.00592 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Inpainting focuses on filling missing or corrupted regions of an image to
blend seamlessly with its surrounding content and style. While conditional
diffusion models have proven effective for text-guided inpainting, we introduce
the novel task of multi-mask inpainting, where multiple regions are
simultaneously inpainted using distinct prompts. Furthermore, we design a
fine-tuning procedure for multimodal LLMs, such as LLaVA, to generate
multi-mask prompts automatically using corrupted images as inputs. These models
can generate helpful and detailed prompt suggestions for filling the masked
regions. The generated prompts are then fed to Stable Diffusion, which is
fine-tuned for the multi-mask inpainting problem using rectified
cross-attention, enforcing prompts onto their designated regions for filling.
Experiments on digitized paintings from WikiArt and the Densely Captioned
Images dataset demonstrate that our pipeline delivers creative and accurate
inpainting results. Our code, data, and trained models are available at
https://cilabuniba.github.io/i-dream-my-painting.
| [
{
"version": "v1",
"created": "Thu, 28 Nov 2024 10:55:09 GMT"
},
{
"version": "v2",
"created": "Fri, 6 Dec 2024 10:58:53 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Fanelli",
"Nicola",
""
],
[
"Vessio",
"Gennaro",
""
],
[
"Castellano",
"Giovanna",
""
]
] | TITLE: I Dream My Painting: Connecting MLLMs and Diffusion Models via Prompt
Generation for Text-Guided Multi-Mask Inpainting
ABSTRACT: Inpainting focuses on filling missing or corrupted regions of an image to
blend seamlessly with its surrounding content and style. While conditional
diffusion models have proven effective for text-guided inpainting, we introduce
the novel task of multi-mask inpainting, where multiple regions are
simultaneously inpainted using distinct prompts. Furthermore, we design a
fine-tuning procedure for multimodal LLMs, such as LLaVA, to generate
multi-mask prompts automatically using corrupted images as inputs. These models
can generate helpful and detailed prompt suggestions for filling the masked
regions. The generated prompts are then fed to Stable Diffusion, which is
fine-tuned for the multi-mask inpainting problem using rectified
cross-attention, enforcing prompts onto their designated regions for filling.
Experiments on digitized paintings from WikiArt and the Densely Captioned
Images dataset demonstrate that our pipeline delivers creative and accurate
inpainting results. Our code, data, and trained models are available at
https://cilabuniba.github.io/i-dream-my-painting.
|
2411.19346 | Mohamed Fazli Imam | Mohamed Fazli Imam, Rufael Fedaku Marew, Jameel Hassan, Mustansar
Fiaz, Alham Fikri Aji, Hisham Cholakkal | CLIP meets DINO for Tuning Zero-Shot Classifier using Unlabeled Image
Collections | null | null | null | null | cs.CV cs.CL cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In the era of foundation models, CLIP has emerged as a powerful tool for
aligning text & visual modalities into a common embedding space. However, the
alignment objective used to train CLIP often results in subpar visual features
for fine-grained tasks. In contrast, SSL-pretrained models like DINO excel at
extracting rich visual features due to their specialized training paradigm.
Yet, these SSL models require an additional supervised linear probing step,
which relies on fully labeled data which is often expensive and difficult to
obtain at scale. In this paper, we propose a label-free prompt-tuning method
that leverages the rich visual features of self-supervised learning models
(DINO) and the broad textual knowledge of large language models (LLMs) to
largely enhance CLIP-based image classification performance using unlabeled
images. Our approach unfolds in three key steps: (1) We generate robust textual
feature embeddings that more accurately represent object classes by leveraging
class-specific descriptions from LLMs, enabling more effective zero-shot
classification compared to CLIP's default name-specific prompts. (2) These
textual embeddings are then used to produce pseudo-labels to train an alignment
module that integrates the complementary strengths of LLM description-based
textual embeddings & DINO's visual features. (3) Finally, we prompt-tune CLIP's
vision encoder through DINO-assisted supervision using the trained alignment
module. This three-step process allows us to harness the best of visual &
textual foundation models, resulting in a powerful and efficient approach that
surpasses state-of-the-art label-free classification methods. Notably, our
framework, NoLA (No Labels Attached), achieves an average absolute gain of 3.6%
over the state-of-the-art LaFTer across 11 diverse image classification
datasets. Our code & models can be found at https://github.com/fazliimam/NoLA.
| [
{
"version": "v1",
"created": "Thu, 28 Nov 2024 19:48:54 GMT"
},
{
"version": "v2",
"created": "Fri, 7 Mar 2025 08:08:18 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 11:09:41 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Imam",
"Mohamed Fazli",
""
],
[
"Marew",
"Rufael Fedaku",
""
],
[
"Hassan",
"Jameel",
""
],
[
"Fiaz",
"Mustansar",
""
],
[
"Aji",
"Alham Fikri",
""
],
[
"Cholakkal",
"Hisham",
""
]
] | TITLE: CLIP meets DINO for Tuning Zero-Shot Classifier using Unlabeled Image
Collections
ABSTRACT: In the era of foundation models, CLIP has emerged as a powerful tool for
aligning text & visual modalities into a common embedding space. However, the
alignment objective used to train CLIP often results in subpar visual features
for fine-grained tasks. In contrast, SSL-pretrained models like DINO excel at
extracting rich visual features due to their specialized training paradigm.
Yet, these SSL models require an additional supervised linear probing step,
which relies on fully labeled data which is often expensive and difficult to
obtain at scale. In this paper, we propose a label-free prompt-tuning method
that leverages the rich visual features of self-supervised learning models
(DINO) and the broad textual knowledge of large language models (LLMs) to
largely enhance CLIP-based image classification performance using unlabeled
images. Our approach unfolds in three key steps: (1) We generate robust textual
feature embeddings that more accurately represent object classes by leveraging
class-specific descriptions from LLMs, enabling more effective zero-shot
classification compared to CLIP's default name-specific prompts. (2) These
textual embeddings are then used to produce pseudo-labels to train an alignment
module that integrates the complementary strengths of LLM description-based
textual embeddings & DINO's visual features. (3) Finally, we prompt-tune CLIP's
vision encoder through DINO-assisted supervision using the trained alignment
module. This three-step process allows us to harness the best of visual &
textual foundation models, resulting in a powerful and efficient approach that
surpasses state-of-the-art label-free classification methods. Notably, our
framework, NoLA (No Labels Attached), achieves an average absolute gain of 3.6%
over the state-of-the-art LaFTer across 11 diverse image classification
datasets. Our code & models can be found at https://github.com/fazliimam/NoLA.
|
2412.01721 | Marc Van Droogenbroeck | Floriane Magera and Thomas Hoyoux and Olivier Barnich and Marc Van
Droogenbroeck | BroadTrack: Broadcast Camera Tracking for Soccer | 12 pages, 4 figures, 3 tables, 60 references | null | 10.1109/wacv61041.2025.00602 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Camera calibration and localization, sometimes simply named camera
calibration, enables many applications in the context of soccer broadcasting,
for instance regarding the interpretation and analysis of the game, or the
insertion of augmented reality graphics for storytelling or refereeing
purposes. To contribute to such applications, the research community has
typically focused on single-view calibration methods, leveraging the
near-omnipresence of soccer field markings in wide-angle broadcast views, but
leaving all temporal aspects, if considered at all, to general-purpose tracking
or filtering techniques. Only a few contributions have been made to leverage
any domain-specific knowledge for this tracking task, and, as a result, there
lacks a truly performant and off-the-shelf camera tracking system tailored for
soccer broadcasting, specifically for elevated tripod-mounted cameras around
the stadium. In this work, we present such a system capable of addressing the
task of soccer broadcast camera tracking efficiently, robustly, and accurately,
outperforming by far the most precise methods of the state-of-the-art. By
combining the available open-source soccer field detectors with carefully
designed camera and tripod models, our tracking system, BroadTrack, halves the
mean reprojection error rate and gains more than 15% in terms of Jaccard index
for camera calibration on the SoccerNet dataset. Furthermore, as the SoccerNet
dataset videos are relatively short (30 seconds), we also present qualitative
results on a 20-minute broadcast clip to showcase the robustness and the
soundness of our system.
| [
{
"version": "v1",
"created": "Mon, 2 Dec 2024 17:10:52 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Magera",
"Floriane",
""
],
[
"Hoyoux",
"Thomas",
""
],
[
"Barnich",
"Olivier",
""
],
[
"Van Droogenbroeck",
"Marc",
""
]
] | TITLE: BroadTrack: Broadcast Camera Tracking for Soccer
ABSTRACT: Camera calibration and localization, sometimes simply named camera
calibration, enables many applications in the context of soccer broadcasting,
for instance regarding the interpretation and analysis of the game, or the
insertion of augmented reality graphics for storytelling or refereeing
purposes. To contribute to such applications, the research community has
typically focused on single-view calibration methods, leveraging the
near-omnipresence of soccer field markings in wide-angle broadcast views, but
leaving all temporal aspects, if considered at all, to general-purpose tracking
or filtering techniques. Only a few contributions have been made to leverage
any domain-specific knowledge for this tracking task, and, as a result, there
lacks a truly performant and off-the-shelf camera tracking system tailored for
soccer broadcasting, specifically for elevated tripod-mounted cameras around
the stadium. In this work, we present such a system capable of addressing the
task of soccer broadcast camera tracking efficiently, robustly, and accurately,
outperforming by far the most precise methods of the state-of-the-art. By
combining the available open-source soccer field detectors with carefully
designed camera and tripod models, our tracking system, BroadTrack, halves the
mean reprojection error rate and gains more than 15% in terms of Jaccard index
for camera calibration on the SoccerNet dataset. Furthermore, as the SoccerNet
dataset videos are relatively short (30 seconds), we also present qualitative
results on a 20-minute broadcast clip to showcase the robustness and the
soundness of our system.
|
2412.08864 | Jiankang Wang | Jiankang Wang, Jianjun Xu, Xiaorui Wang, Yuxin Wang, Mengting Xing,
Shancheng Fang, Zhineng Chen, Hongtao Xie, Yongdong Zhang | A Graph-Based Synthetic Data Pipeline for Scaling High-Quality Reasoning
Instructions | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Synthesizing high-quality reasoning data for continual training has been
proven to be effective in enhancing the performance of Large Language Models
(LLMs). However, previous synthetic approaches struggle to easily scale up data
and incur high costs in the pursuit of high quality. In this paper, we propose
the Graph-based Synthetic Data Pipeline (GSDP), an economical and scalable
framework for high-quality reasoning data synthesis. Inspired by knowledge
graphs, we extracted knowledge points from seed data and constructed a
knowledge point relationships graph to explore their interconnections. By
exploring the implicit relationships among knowledge, our method achieves
$\times$255 data expansion. Furthermore, GSDP led by open-source models,
achieves synthesis quality comparable to GPT-4-0613 while maintaining
$\times$100 lower costs. To tackle the most challenging mathematical reasoning
task, we present the GSDP-MATH dataset comprising over 1.91 million pairs of
math problems and answers. After fine-tuning on GSDP-MATH, GSDP-7B based on
Mistral-7B achieves 37.7% accuracy on MATH and 78.4% on GSM8K, demonstrating
the effectiveness of our method. The dataset and models will be released in
https://github.com/Jayce1kk/GSDP.
| [
{
"version": "v1",
"created": "Thu, 12 Dec 2024 01:52:25 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 10:47:53 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Wang",
"Jiankang",
""
],
[
"Xu",
"Jianjun",
""
],
[
"Wang",
"Xiaorui",
""
],
[
"Wang",
"Yuxin",
""
],
[
"Xing",
"Mengting",
""
],
[
"Fang",
"Shancheng",
""
],
[
"Chen",
"Zhineng",
""
],
[
"Xie",
"Hongtao",
""
],
[
"Zhang",
"Yongdong",
""
]
] | TITLE: A Graph-Based Synthetic Data Pipeline for Scaling High-Quality Reasoning
Instructions
ABSTRACT: Synthesizing high-quality reasoning data for continual training has been
proven to be effective in enhancing the performance of Large Language Models
(LLMs). However, previous synthetic approaches struggle to easily scale up data
and incur high costs in the pursuit of high quality. In this paper, we propose
the Graph-based Synthetic Data Pipeline (GSDP), an economical and scalable
framework for high-quality reasoning data synthesis. Inspired by knowledge
graphs, we extracted knowledge points from seed data and constructed a
knowledge point relationships graph to explore their interconnections. By
exploring the implicit relationships among knowledge, our method achieves
$\times$255 data expansion. Furthermore, GSDP led by open-source models,
achieves synthesis quality comparable to GPT-4-0613 while maintaining
$\times$100 lower costs. To tackle the most challenging mathematical reasoning
task, we present the GSDP-MATH dataset comprising over 1.91 million pairs of
math problems and answers. After fine-tuning on GSDP-MATH, GSDP-7B based on
Mistral-7B achieves 37.7% accuracy on MATH and 78.4% on GSM8K, demonstrating
the effectiveness of our method. The dataset and models will be released in
https://github.com/Jayce1kk/GSDP.
|
2412.08912 | Ali Mollaahmadi Dehaghi | Ali Mollaahmadi Dehaghi, Reza Razavi, Mohammad Moshirpour | Reversing the Damage: A QP-Aware Transformer-Diffusion Approach for 8K
Video Restoration under Codec Compression | 12 pages, 8 figures | null | 10.1109/WACV61041.2025.00130 | null | cs.CV cs.MM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we introduce DiQP; a novel Transformer-Diffusion model for
restoring 8K video quality degraded by codec compression. To the best of our
knowledge, our model is the first to consider restoring the artifacts
introduced by various codecs (AV1, HEVC) by Denoising Diffusion without
considering additional noise. This approach allows us to model the complex,
non-Gaussian nature of compression artifacts, effectively learning to reverse
the degradation. Our architecture combines the power of Transformers to capture
long-range dependencies with an enhanced windowed mechanism that preserves
spatiotemporal context within groups of pixels across frames. To further
enhance restoration, the model incorporates auxiliary "Look Ahead" and "Look
Around" modules, providing both future and surrounding frame information to aid
in reconstructing fine details and enhancing overall visual quality. Extensive
experiments on different datasets demonstrate that our model outperforms
state-of-the-art methods, particularly for high-resolution videos such as 4K
and 8K, showcasing its effectiveness in restoring perceptually pleasing videos
from highly compressed sources.
| [
{
"version": "v1",
"created": "Thu, 12 Dec 2024 03:49:22 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Dehaghi",
"Ali Mollaahmadi",
""
],
[
"Razavi",
"Reza",
""
],
[
"Moshirpour",
"Mohammad",
""
]
] | TITLE: Reversing the Damage: A QP-Aware Transformer-Diffusion Approach for 8K
Video Restoration under Codec Compression
ABSTRACT: In this paper, we introduce DiQP; a novel Transformer-Diffusion model for
restoring 8K video quality degraded by codec compression. To the best of our
knowledge, our model is the first to consider restoring the artifacts
introduced by various codecs (AV1, HEVC) by Denoising Diffusion without
considering additional noise. This approach allows us to model the complex,
non-Gaussian nature of compression artifacts, effectively learning to reverse
the degradation. Our architecture combines the power of Transformers to capture
long-range dependencies with an enhanced windowed mechanism that preserves
spatiotemporal context within groups of pixels across frames. To further
enhance restoration, the model incorporates auxiliary "Look Ahead" and "Look
Around" modules, providing both future and surrounding frame information to aid
in reconstructing fine details and enhancing overall visual quality. Extensive
experiments on different datasets demonstrate that our model outperforms
state-of-the-art methods, particularly for high-resolution videos such as 4K
and 8K, showcasing its effectiveness in restoring perceptually pleasing videos
from highly compressed sources.
|
2412.14602 | Yuxuan Liang | Yuxuan Liang, Wentao Zhang, Zeang Sheng, Ling Yang, Quanqing Xu,
Jiawei Jiang, Yunhai Tong, Bin Cui | Towards Scalable and Deep Graph Neural Networks via Noise Masking | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, Graph Neural Networks (GNNs) have achieved remarkable
success in many graph mining tasks. However, scaling them to large graphs is
challenging due to the high computational and storage costs of repeated feature
propagation and non-linear transformation during training. One commonly
employed approach to address this challenge is model-simplification, which only
executes the Propagation (P) once in the pre-processing, and Combine (C) these
receptive fields in different ways and then feed them into a simple model for
better performance. Despite their high predictive performance and scalability,
these methods still face two limitations. First, existing approaches mainly
focus on exploring different C methods from the model perspective, neglecting
the crucial problem of performance degradation with increasing P depth from the
data-centric perspective, known as the over-smoothing problem. Second,
pre-processing overhead takes up most of the end-to-end processing time,
especially for large-scale graphs. To address these limitations, we present
random walk with noise masking (RMask), a plug-and-play module compatible with
the existing model-simplification works. This module enables the exploration of
deeper GNNs while preserving their scalability. Unlike the previous
model-simplification works, we focus on continuous P and found that the noise
existing inside each P is the cause of the over-smoothing issue, and use the
efficient masking mechanism to eliminate them. Experimental results on six
real-world datasets demonstrate that model-simplification works equipped with
RMask yield superior performance compared to their original version and can
make a good trade-off between accuracy and efficiency.
| [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 07:48:14 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 02:16:19 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Liang",
"Yuxuan",
""
],
[
"Zhang",
"Wentao",
""
],
[
"Sheng",
"Zeang",
""
],
[
"Yang",
"Ling",
""
],
[
"Xu",
"Quanqing",
""
],
[
"Jiang",
"Jiawei",
""
],
[
"Tong",
"Yunhai",
""
],
[
"Cui",
"Bin",
""
]
] | TITLE: Towards Scalable and Deep Graph Neural Networks via Noise Masking
ABSTRACT: In recent years, Graph Neural Networks (GNNs) have achieved remarkable
success in many graph mining tasks. However, scaling them to large graphs is
challenging due to the high computational and storage costs of repeated feature
propagation and non-linear transformation during training. One commonly
employed approach to address this challenge is model-simplification, which only
executes the Propagation (P) once in the pre-processing, and Combine (C) these
receptive fields in different ways and then feed them into a simple model for
better performance. Despite their high predictive performance and scalability,
these methods still face two limitations. First, existing approaches mainly
focus on exploring different C methods from the model perspective, neglecting
the crucial problem of performance degradation with increasing P depth from the
data-centric perspective, known as the over-smoothing problem. Second,
pre-processing overhead takes up most of the end-to-end processing time,
especially for large-scale graphs. To address these limitations, we present
random walk with noise masking (RMask), a plug-and-play module compatible with
the existing model-simplification works. This module enables the exploration of
deeper GNNs while preserving their scalability. Unlike the previous
model-simplification works, we focus on continuous P and found that the noise
existing inside each P is the cause of the over-smoothing issue, and use the
efficient masking mechanism to eliminate them. Experimental results on six
real-world datasets demonstrate that model-simplification works equipped with
RMask yield superior performance compared to their original version and can
make a good trade-off between accuracy and efficiency.
|
2412.14719 | Kun Li | Kun Li, Dan Guo, Guoliang Chen, Chunxiao Fan, Jingyuan Xu, Zhiliang
Wu, Hehe Fan, Meng Wang | Prototypical Calibrating Ambiguous Samples for Micro-Action Recognition | Fix typos; Accepted by AAAI 2025 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Micro-Action Recognition (MAR) has gained increasing attention due to its
crucial role as a form of non-verbal communication in social interactions, with
promising potential for applications in human communication and emotion
analysis. However, current approaches often overlook the inherent ambiguity in
micro-actions, which arises from the wide category range and subtle visual
differences between categories. This oversight hampers the accuracy of
micro-action recognition. In this paper, we propose a novel Prototypical
Calibrating Ambiguous Network (PCAN) to unleash and mitigate the ambiguity of
MAR. Firstly, we employ a hierarchical action-tree to identify the ambiguous
sample, categorizing them into distinct sets of ambiguous samples of false
negatives and false positives, considering both body- and action-level
categories. Secondly, we implement an ambiguous contrastive refinement module
to calibrate these ambiguous samples by regulating the distance between
ambiguous samples and their corresponding prototypes. This calibration process
aims to pull false negative (FN) samples closer to their respective prototypes
and push false positive (FP) samples apart from their affiliated prototypes. In
addition, we propose a new prototypical diversity amplification loss to
strengthen the model's capacity by amplifying the differences between different
prototypes. Finally, we propose a prototype-guided rectification to rectify
prediction by incorporating the representability of prototypes. Extensive
experiments conducted on the benchmark dataset demonstrate the superior
performance of our method compared to existing approaches. The code is
available at https://github.com/kunli-cs/PCAN.
| [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 10:41:24 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 05:13:15 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Li",
"Kun",
""
],
[
"Guo",
"Dan",
""
],
[
"Chen",
"Guoliang",
""
],
[
"Fan",
"Chunxiao",
""
],
[
"Xu",
"Jingyuan",
""
],
[
"Wu",
"Zhiliang",
""
],
[
"Fan",
"Hehe",
""
],
[
"Wang",
"Meng",
""
]
] | TITLE: Prototypical Calibrating Ambiguous Samples for Micro-Action Recognition
ABSTRACT: Micro-Action Recognition (MAR) has gained increasing attention due to its
crucial role as a form of non-verbal communication in social interactions, with
promising potential for applications in human communication and emotion
analysis. However, current approaches often overlook the inherent ambiguity in
micro-actions, which arises from the wide category range and subtle visual
differences between categories. This oversight hampers the accuracy of
micro-action recognition. In this paper, we propose a novel Prototypical
Calibrating Ambiguous Network (PCAN) to unleash and mitigate the ambiguity of
MAR. Firstly, we employ a hierarchical action-tree to identify the ambiguous
sample, categorizing them into distinct sets of ambiguous samples of false
negatives and false positives, considering both body- and action-level
categories. Secondly, we implement an ambiguous contrastive refinement module
to calibrate these ambiguous samples by regulating the distance between
ambiguous samples and their corresponding prototypes. This calibration process
aims to pull false negative (FN) samples closer to their respective prototypes
and push false positive (FP) samples apart from their affiliated prototypes. In
addition, we propose a new prototypical diversity amplification loss to
strengthen the model's capacity by amplifying the differences between different
prototypes. Finally, we propose a prototype-guided rectification to rectify
prediction by incorporating the representability of prototypes. Extensive
experiments conducted on the benchmark dataset demonstrate the superior
performance of our method compared to existing approaches. The code is
available at https://github.com/kunli-cs/PCAN.
|
2412.17534 | Ege Yi\u{g}it \c{C}elik | Ege Yi\u{g}it \c{C}elik and Selma Tekir | CiteBART: Learning to Generate Citations for Local Citation
Recommendation | 17 pages, 2 figures, 10 tables | null | null | null | cs.IR cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Local citation recommendation (LCR) suggests a set of papers for a citation
placeholder within a given context. The task has evolved as generative
approaches have become more promising than the traditional pre-fetch and
re-rank-based state-of-the-art approaches. This paper introduces
citation-specific pre-training within an encoder-decoder architecture, where
author-date citation tokens are masked to learn to reconstruct them to fulfill
LCR. There are two variants for this pre-training. In the local context-only
base scheme (CiteBART-Base), the citation token in a local context is masked to
learn to predict the citation. The global version (CiteBART-Global) extends the
local context with the citing paper's title and abstract to enrich the learning
signal. CiteBART-Global achieves state-of-the-art performance on LCR benchmarks
except for the FullTextPeerRead dataset, which is quite small to see the
advantage of generative pre-training. The effect is significant in the larger
benchmarks, e.g., Refseer and ArXiv., with the Refseer benchmark-trained model
emerging as the best-performing model. We perform comprehensive experiments,
including an ablation study, a qualitative analysis, and a taxonomy of
hallucinations with detailed statistics. Our analyses confirm that
CiteBART-Global has a cross-dataset generalization capability; the macro
hallucination rate (MaHR) at the top-3 predictions is 4\%, and when the
ground-truth is in the top-k prediction list, the hallucination tendency in the
other predictions drops significantly.
| [
{
"version": "v1",
"created": "Mon, 23 Dec 2024 12:58:30 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Apr 2025 20:23:16 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Çelik",
"Ege Yiğit",
""
],
[
"Tekir",
"Selma",
""
]
] | TITLE: CiteBART: Learning to Generate Citations for Local Citation
Recommendation
ABSTRACT: Local citation recommendation (LCR) suggests a set of papers for a citation
placeholder within a given context. The task has evolved as generative
approaches have become more promising than the traditional pre-fetch and
re-rank-based state-of-the-art approaches. This paper introduces
citation-specific pre-training within an encoder-decoder architecture, where
author-date citation tokens are masked to learn to reconstruct them to fulfill
LCR. There are two variants for this pre-training. In the local context-only
base scheme (CiteBART-Base), the citation token in a local context is masked to
learn to predict the citation. The global version (CiteBART-Global) extends the
local context with the citing paper's title and abstract to enrich the learning
signal. CiteBART-Global achieves state-of-the-art performance on LCR benchmarks
except for the FullTextPeerRead dataset, which is quite small to see the
advantage of generative pre-training. The effect is significant in the larger
benchmarks, e.g., Refseer and ArXiv., with the Refseer benchmark-trained model
emerging as the best-performing model. We perform comprehensive experiments,
including an ablation study, a qualitative analysis, and a taxonomy of
hallucinations with detailed statistics. Our analyses confirm that
CiteBART-Global has a cross-dataset generalization capability; the macro
hallucination rate (MaHR) at the top-3 predictions is 4\%, and when the
ground-truth is in the top-k prediction list, the hallucination tendency in the
other predictions drops significantly.
|
2412.20439 | Wangyu Wu | Wangyu Wu, Xianglin Qiu, Siqi Song, Zhenhong Chen, Xiaowei Huang, Fei
Ma, Jimin Xiao | Image Augmentation Agent for Weakly Supervised Semantic Segmentation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Weakly-supervised semantic segmentation (WSSS) has achieved remarkable
progress using only image-level labels. However, most existing WSSS methods
focus on designing new network structures and loss functions to generate more
accurate dense labels, overlooking the limitations imposed by fixed datasets,
which can constrain performance improvements. We argue that more diverse
trainable images provides WSSS richer information and help model understand
more comprehensive semantic pattern. Therefore in this paper, we introduce a
novel approach called Image Augmentation Agent (IAA) which shows that it is
possible to enhance WSSS from data generation perspective. IAA mainly design an
augmentation agent that leverages large language models (LLMs) and diffusion
models to automatically generate additional images for WSSS. In practice, to
address the instability in prompt generation by LLMs, we develop a prompt
self-refinement mechanism. It allow LLMs to re-evaluate the rationality of
generated prompts to produce more coherent prompts. Additionally, we insert an
online filter into diffusion generation process to dynamically ensure the
quality and balance of generated images. Experimental results show that our
method significantly surpasses state-of-the-art WSSS approaches on the PASCAL
VOC 2012 and MS COCO 2014 datasets.
| [
{
"version": "v1",
"created": "Sun, 29 Dec 2024 11:32:55 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 08:36:11 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Wu",
"Wangyu",
""
],
[
"Qiu",
"Xianglin",
""
],
[
"Song",
"Siqi",
""
],
[
"Chen",
"Zhenhong",
""
],
[
"Huang",
"Xiaowei",
""
],
[
"Ma",
"Fei",
""
],
[
"Xiao",
"Jimin",
""
]
] | TITLE: Image Augmentation Agent for Weakly Supervised Semantic Segmentation
ABSTRACT: Weakly-supervised semantic segmentation (WSSS) has achieved remarkable
progress using only image-level labels. However, most existing WSSS methods
focus on designing new network structures and loss functions to generate more
accurate dense labels, overlooking the limitations imposed by fixed datasets,
which can constrain performance improvements. We argue that more diverse
trainable images provides WSSS richer information and help model understand
more comprehensive semantic pattern. Therefore in this paper, we introduce a
novel approach called Image Augmentation Agent (IAA) which shows that it is
possible to enhance WSSS from data generation perspective. IAA mainly design an
augmentation agent that leverages large language models (LLMs) and diffusion
models to automatically generate additional images for WSSS. In practice, to
address the instability in prompt generation by LLMs, we develop a prompt
self-refinement mechanism. It allow LLMs to re-evaluate the rationality of
generated prompts to produce more coherent prompts. Additionally, we insert an
online filter into diffusion generation process to dynamically ensure the
quality and balance of generated images. Experimental results show that our
method significantly surpasses state-of-the-art WSSS approaches on the PASCAL
VOC 2012 and MS COCO 2014 datasets.
|
2412.21037 | Soujanya Poria | Chia-Yu Hung, Navonil Majumder, Zhifeng Kong, Ambuj Mehrish, Amir Ali
Bagherzadeh, Chuan Li, Rafael Valle, Bryan Catanzaro, Soujanya Poria | TangoFlux: Super Fast and Faithful Text to Audio Generation with Flow
Matching and Clap-Ranked Preference Optimization | https://tangoflux.github.io/ | null | null | null | cs.SD cs.AI cs.CL eess.AS | http://creativecommons.org/licenses/by-sa/4.0/ | We introduce TangoFlux, an efficient Text-to-Audio (TTA) generative model
with 515M parameters, capable of generating up to 30 seconds of 44.1kHz audio
in just 3.7 seconds on a single A40 GPU. A key challenge in aligning TTA models
lies in the difficulty of creating preference pairs, as TTA lacks structured
mechanisms like verifiable rewards or gold-standard answers available for Large
Language Models (LLMs). To address this, we propose CLAP-Ranked Preference
Optimization (CRPO), a novel framework that iteratively generates and optimizes
preference data to enhance TTA alignment. We demonstrate that the audio
preference dataset generated using CRPO outperforms existing alternatives. With
this framework, TangoFlux achieves state-of-the-art performance across both
objective and subjective benchmarks. We open source all code and models to
support further research in TTA generation.
| [
{
"version": "v1",
"created": "Mon, 30 Dec 2024 16:02:44 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 05:01:32 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Hung",
"Chia-Yu",
""
],
[
"Majumder",
"Navonil",
""
],
[
"Kong",
"Zhifeng",
""
],
[
"Mehrish",
"Ambuj",
""
],
[
"Bagherzadeh",
"Amir Ali",
""
],
[
"Li",
"Chuan",
""
],
[
"Valle",
"Rafael",
""
],
[
"Catanzaro",
"Bryan",
""
],
[
"Poria",
"Soujanya",
""
]
] | TITLE: TangoFlux: Super Fast and Faithful Text to Audio Generation with Flow
Matching and Clap-Ranked Preference Optimization
ABSTRACT: We introduce TangoFlux, an efficient Text-to-Audio (TTA) generative model
with 515M parameters, capable of generating up to 30 seconds of 44.1kHz audio
in just 3.7 seconds on a single A40 GPU. A key challenge in aligning TTA models
lies in the difficulty of creating preference pairs, as TTA lacks structured
mechanisms like verifiable rewards or gold-standard answers available for Large
Language Models (LLMs). To address this, we propose CLAP-Ranked Preference
Optimization (CRPO), a novel framework that iteratively generates and optimizes
preference data to enhance TTA alignment. We demonstrate that the audio
preference dataset generated using CRPO outperforms existing alternatives. With
this framework, TangoFlux achieves state-of-the-art performance across both
objective and subjective benchmarks. We open source all code and models to
support further research in TTA generation.
|
2501.05449 | MD Arafat Alam Khandaker Arafat | Md. Arafat Alam Khandaker, Ziyan Shirin Raha, Shifat Islam, Tashreef
Muhammad | Explainable AI-Enhanced Deep Learning for Pumpkin Leaf Disease
Detection: A Comparative Analysis of CNN Architectures | Accepted in 2024 27th International Conference on Computer and
Information Technology (ICCIT) | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Pumpkin leaf diseases are significant threats to agricultural productivity,
requiring a timely and precise diagnosis for effective management. Traditional
identification methods are laborious and susceptible to human error,
emphasizing the necessity for automated solutions. This study employs on the
"Pumpkin Leaf Disease Dataset", that comprises of 2000 high-resolution images
separated into five categories. Downy mildew, powdery mildew, mosaic disease,
bacterial leaf spot, and healthy leaves. The dataset was rigorously assembled
from several agricultural fields to ensure a strong representation for model
training. We explored many proficient deep learning architectures, including
DenseNet201, DenseNet121, DenseNet169, Xception, ResNet50, ResNet101 and
InceptionResNetV2, and observed that ResNet50 performed most effectively, with
an accuracy of 90.5% and comparable precision, recall, and F1-Score. We used
Explainable AI (XAI) approaches like Grad-CAM, Grad-CAM++, Score-CAM, and
Layer-CAM to provide meaningful representations of model decision-making
processes, which improved understanding and trust in automated disease
diagnostics. These findings demonstrate ResNet50's potential to revolutionize
pumpkin leaf disease detection, allowing for earlier and more accurate
treatments.
| [
{
"version": "v1",
"created": "Thu, 9 Jan 2025 18:59:35 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 17:35:24 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Khandaker",
"Md. Arafat Alam",
""
],
[
"Raha",
"Ziyan Shirin",
""
],
[
"Islam",
"Shifat",
""
],
[
"Muhammad",
"Tashreef",
""
]
] | TITLE: Explainable AI-Enhanced Deep Learning for Pumpkin Leaf Disease
Detection: A Comparative Analysis of CNN Architectures
ABSTRACT: Pumpkin leaf diseases are significant threats to agricultural productivity,
requiring a timely and precise diagnosis for effective management. Traditional
identification methods are laborious and susceptible to human error,
emphasizing the necessity for automated solutions. This study employs on the
"Pumpkin Leaf Disease Dataset", that comprises of 2000 high-resolution images
separated into five categories. Downy mildew, powdery mildew, mosaic disease,
bacterial leaf spot, and healthy leaves. The dataset was rigorously assembled
from several agricultural fields to ensure a strong representation for model
training. We explored many proficient deep learning architectures, including
DenseNet201, DenseNet121, DenseNet169, Xception, ResNet50, ResNet101 and
InceptionResNetV2, and observed that ResNet50 performed most effectively, with
an accuracy of 90.5% and comparable precision, recall, and F1-Score. We used
Explainable AI (XAI) approaches like Grad-CAM, Grad-CAM++, Score-CAM, and
Layer-CAM to provide meaningful representations of model decision-making
processes, which improved understanding and trust in automated disease
diagnostics. These findings demonstrate ResNet50's potential to revolutionize
pumpkin leaf disease detection, allowing for earlier and more accurate
treatments.
|
2501.06465 | Ye Chen | Ye Chen, Dongdong Huang, Haoyun Xu, Cong Fu, Lin Sheng, Qingli Zhou,
Yuqiang Shen, Kai Wang | MedCT: A Clinical Terminology Graph for Generative AI Applications in
Healthcare | Accepted into ICCS 2025 and published in Springer's LNCS Series | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | We introduce the world's first clinical terminology for the Chinese
healthcare community, namely MedCT, accompanied by a clinical foundation model
MedBERT and an entity linking model MedLink. The MedCT system enables
standardized and programmable representation of Chinese clinical data,
successively stimulating the development of new medicines, treatment pathways,
and better patient outcomes for the populous Chinese community. Moreover, the
MedCT knowledge graph provides a principled mechanism to minimize the
hallucination problem of large language models (LLMs), therefore achieving
significant levels of accuracy and safety in LLM-based clinical applications.
By leveraging the LLMs' emergent capabilities of generativeness and
expressiveness, we were able to rapidly built a production-quality terminology
system and deployed to real-world clinical field within three months, while
classical terminologies like SNOMED CT have gone through more than twenty years
development. Our experiments show that the MedCT system achieves
state-of-the-art (SOTA) performance in semantic matching and entity linking
tasks, not only for Chinese but also for English. We also conducted a
longitudinal field experiment by applying MedCT and LLMs in a representative
spectrum of clinical tasks, including electronic health record (EHR)
auto-generation and medical document search for diagnostic decision making. Our
study shows a multitude of values of MedCT for clinical workflows and patient
outcomes, especially in the new genre of clinical LLM applications. We present
our approach in sufficient engineering detail, such that implementing a
clinical terminology for other non-English societies should be readily
reproducible. We openly release our terminology, models and algorithms, along
with real-world clinical datasets for the development.
| [
{
"version": "v1",
"created": "Sat, 11 Jan 2025 07:35:51 GMT"
},
{
"version": "v2",
"created": "Tue, 21 Jan 2025 01:56:11 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 07:29:10 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Chen",
"Ye",
""
],
[
"Huang",
"Dongdong",
""
],
[
"Xu",
"Haoyun",
""
],
[
"Fu",
"Cong",
""
],
[
"Sheng",
"Lin",
""
],
[
"Zhou",
"Qingli",
""
],
[
"Shen",
"Yuqiang",
""
],
[
"Wang",
"Kai",
""
]
] | TITLE: MedCT: A Clinical Terminology Graph for Generative AI Applications in
Healthcare
ABSTRACT: We introduce the world's first clinical terminology for the Chinese
healthcare community, namely MedCT, accompanied by a clinical foundation model
MedBERT and an entity linking model MedLink. The MedCT system enables
standardized and programmable representation of Chinese clinical data,
successively stimulating the development of new medicines, treatment pathways,
and better patient outcomes for the populous Chinese community. Moreover, the
MedCT knowledge graph provides a principled mechanism to minimize the
hallucination problem of large language models (LLMs), therefore achieving
significant levels of accuracy and safety in LLM-based clinical applications.
By leveraging the LLMs' emergent capabilities of generativeness and
expressiveness, we were able to rapidly built a production-quality terminology
system and deployed to real-world clinical field within three months, while
classical terminologies like SNOMED CT have gone through more than twenty years
development. Our experiments show that the MedCT system achieves
state-of-the-art (SOTA) performance in semantic matching and entity linking
tasks, not only for Chinese but also for English. We also conducted a
longitudinal field experiment by applying MedCT and LLMs in a representative
spectrum of clinical tasks, including electronic health record (EHR)
auto-generation and medical document search for diagnostic decision making. Our
study shows a multitude of values of MedCT for clinical workflows and patient
outcomes, especially in the new genre of clinical LLM applications. We present
our approach in sufficient engineering detail, such that implementing a
clinical terminology for other non-English societies should be readily
reproducible. We openly release our terminology, models and algorithms, along
with real-world clinical datasets for the development.
|
2501.07824 | Joonho Ko | Joonho Ko, Jinheon Baek, Sung Ju Hwang | Real-time Verification and Refinement of Language Model Text Generation | null | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have shown remarkable performance across a wide
range of natural language tasks. However, a critical challenge remains in that
they sometimes generate factually incorrect answers. To address this, while
many previous work has focused on identifying errors in their generation and
further refining them, they are slow in deployment since they are designed to
verify the response from LLMs only after their entire generation (from the
first to last tokens) is done. Further, we observe that once LLMs generate
incorrect tokens early on, there is a higher likelihood that subsequent tokens
will also be factually incorrect. To this end, in this work, we propose
Streaming-VR (Streaming Verification and Refinement), a novel approach designed
to enhance the efficiency of verification and refinement of LLM outputs.
Specifically, the proposed Streaming-VR enables on-the-fly verification and
correction of tokens as they are being generated, similar to a streaming
process, ensuring that each subset of tokens is checked and refined in
real-time by another LLM as the LLM constructs its response. Through
comprehensive evaluations on multiple datasets, we demonstrate that our
approach not only enhances the factual accuracy of LLMs, but also offers a more
efficient solution compared to prior refinement methods.
| [
{
"version": "v1",
"created": "Tue, 14 Jan 2025 03:59:48 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Feb 2025 13:26:52 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 06:39:35 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Ko",
"Joonho",
""
],
[
"Baek",
"Jinheon",
""
],
[
"Hwang",
"Sung Ju",
""
]
] | TITLE: Real-time Verification and Refinement of Language Model Text Generation
ABSTRACT: Large language models (LLMs) have shown remarkable performance across a wide
range of natural language tasks. However, a critical challenge remains in that
they sometimes generate factually incorrect answers. To address this, while
many previous work has focused on identifying errors in their generation and
further refining them, they are slow in deployment since they are designed to
verify the response from LLMs only after their entire generation (from the
first to last tokens) is done. Further, we observe that once LLMs generate
incorrect tokens early on, there is a higher likelihood that subsequent tokens
will also be factually incorrect. To this end, in this work, we propose
Streaming-VR (Streaming Verification and Refinement), a novel approach designed
to enhance the efficiency of verification and refinement of LLM outputs.
Specifically, the proposed Streaming-VR enables on-the-fly verification and
correction of tokens as they are being generated, similar to a streaming
process, ensuring that each subset of tokens is checked and refined in
real-time by another LLM as the LLM constructs its response. Through
comprehensive evaluations on multiple datasets, we demonstrate that our
approach not only enhances the factual accuracy of LLMs, but also offers a more
efficient solution compared to prior refinement methods.
|
2501.14174 | Junyeob Baek | Junyeob Baek, Yi-Fu Wu, Gautam Singh, Sungjin Ahn | Dreamweaver: Learning Compositional World Models from Pixels | null | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Humans have an innate ability to decompose their perceptions of the world
into objects and their attributes, such as colors, shapes, and movement
patterns. This cognitive process enables us to imagine novel futures by
recombining familiar concepts. However, replicating this ability in artificial
intelligence systems has proven challenging, particularly when it comes to
modeling videos into compositional concepts and generating unseen, recomposed
futures without relying on auxiliary data, such as text, masks, or bounding
boxes. In this paper, we propose Dreamweaver, a neural architecture designed to
discover hierarchical and compositional representations from raw videos and
generate compositional future simulations. Our approach leverages a novel
Recurrent Block-Slot Unit (RBSU) to decompose videos into their constituent
objects and attributes. In addition, Dreamweaver uses a multi-future-frame
prediction objective to capture disentangled representations for dynamic
concepts more effectively as well as static concepts. In experiments, we
demonstrate our model outperforms current state-of-the-art baselines for world
modeling when evaluated under the DCI framework across multiple datasets.
Furthermore, we show how the modularized concept representations of our model
enable compositional imagination, allowing the generation of novel videos by
recombining attributes from previously seen objects.
cun-bjy.github.io/dreamweaver-website
| [
{
"version": "v1",
"created": "Fri, 24 Jan 2025 01:50:19 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Feb 2025 08:16:03 GMT"
},
{
"version": "v3",
"created": "Thu, 27 Feb 2025 16:09:15 GMT"
},
{
"version": "v4",
"created": "Fri, 28 Feb 2025 08:12:21 GMT"
},
{
"version": "v5",
"created": "Thu, 10 Apr 2025 13:12:34 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Baek",
"Junyeob",
""
],
[
"Wu",
"Yi-Fu",
""
],
[
"Singh",
"Gautam",
""
],
[
"Ahn",
"Sungjin",
""
]
] | TITLE: Dreamweaver: Learning Compositional World Models from Pixels
ABSTRACT: Humans have an innate ability to decompose their perceptions of the world
into objects and their attributes, such as colors, shapes, and movement
patterns. This cognitive process enables us to imagine novel futures by
recombining familiar concepts. However, replicating this ability in artificial
intelligence systems has proven challenging, particularly when it comes to
modeling videos into compositional concepts and generating unseen, recomposed
futures without relying on auxiliary data, such as text, masks, or bounding
boxes. In this paper, we propose Dreamweaver, a neural architecture designed to
discover hierarchical and compositional representations from raw videos and
generate compositional future simulations. Our approach leverages a novel
Recurrent Block-Slot Unit (RBSU) to decompose videos into their constituent
objects and attributes. In addition, Dreamweaver uses a multi-future-frame
prediction objective to capture disentangled representations for dynamic
concepts more effectively as well as static concepts. In experiments, we
demonstrate our model outperforms current state-of-the-art baselines for world
modeling when evaluated under the DCI framework across multiple datasets.
Furthermore, we show how the modularized concept representations of our model
enable compositional imagination, allowing the generation of novel videos by
recombining attributes from previously seen objects.
cun-bjy.github.io/dreamweaver-website
|
2501.15293 | Sajjad Saleem | Rubab Hafeez, Sadia Waheed, Syeda Aleena Naqvi, Fahad Maqbool, Amna
Sarwar, Sajjad Saleem, Muhammad Imran Sharif, Kamran Siddique, Zahid Akhtar | Deep Learning in Early Alzheimer's disease's Detection: A Comprehensive
Survey of Classification, Segmentation, and Feature Extraction Methods | 22 pages | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Alzheimers disease is a deadly neurological condition, impairing important
memory and brain functions. Alzheimers disease promotes brain shrinkage,
ultimately leading to dementia. Dementia diagnosis typically takes 2.8 to 4.4
years after the first clinical indication. Advancements in computing and
information technology have led to many techniques of studying Alzheimers
disease. Early identification and therapy are crucial for preventing Alzheimers
disease, as early-onset dementia hits people before the age of 65, while
late-onset dementia occurs after this age. According to the 2015 World
Alzheimers disease Report, there are 46.8 million individuals worldwide
suffering from dementia, with an anticipated 74.7 million more by 2030 and
131.5 million by 2050. Deep Learning has outperformed conventional Machine
Learning techniques by identifying intricate structures in high-dimensional
data. Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN),
have achieved an accuracy of up to 96.0% for Alzheimers disease classification,
and 84.2% for mild cognitive impairment (MCI) conversion prediction. There have
been few literature surveys available on applying ML to predict dementia,
lacking in congenital observations. However, this survey has focused on a
specific data channel for dementia detection. This study evaluated Deep
Learning algorithms for early Alzheimers disease detection, using openly
accessible datasets, feature segmentation, and classification methods. This
article also has identified research gaps and limits in detecting Alzheimers
disease, which can inform future research.
| [
{
"version": "v1",
"created": "Sat, 25 Jan 2025 18:00:17 GMT"
},
{
"version": "v2",
"created": "Wed, 29 Jan 2025 10:30:35 GMT"
},
{
"version": "v3",
"created": "Wed, 9 Apr 2025 22:39:50 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Hafeez",
"Rubab",
""
],
[
"Waheed",
"Sadia",
""
],
[
"Naqvi",
"Syeda Aleena",
""
],
[
"Maqbool",
"Fahad",
""
],
[
"Sarwar",
"Amna",
""
],
[
"Saleem",
"Sajjad",
""
],
[
"Sharif",
"Muhammad Imran",
""
],
[
"Siddique",
"Kamran",
""
],
[
"Akhtar",
"Zahid",
""
]
] | TITLE: Deep Learning in Early Alzheimer's disease's Detection: A Comprehensive
Survey of Classification, Segmentation, and Feature Extraction Methods
ABSTRACT: Alzheimers disease is a deadly neurological condition, impairing important
memory and brain functions. Alzheimers disease promotes brain shrinkage,
ultimately leading to dementia. Dementia diagnosis typically takes 2.8 to 4.4
years after the first clinical indication. Advancements in computing and
information technology have led to many techniques of studying Alzheimers
disease. Early identification and therapy are crucial for preventing Alzheimers
disease, as early-onset dementia hits people before the age of 65, while
late-onset dementia occurs after this age. According to the 2015 World
Alzheimers disease Report, there are 46.8 million individuals worldwide
suffering from dementia, with an anticipated 74.7 million more by 2030 and
131.5 million by 2050. Deep Learning has outperformed conventional Machine
Learning techniques by identifying intricate structures in high-dimensional
data. Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN),
have achieved an accuracy of up to 96.0% for Alzheimers disease classification,
and 84.2% for mild cognitive impairment (MCI) conversion prediction. There have
been few literature surveys available on applying ML to predict dementia,
lacking in congenital observations. However, this survey has focused on a
specific data channel for dementia detection. This study evaluated Deep
Learning algorithms for early Alzheimers disease detection, using openly
accessible datasets, feature segmentation, and classification methods. This
article also has identified research gaps and limits in detecting Alzheimers
disease, which can inform future research.
|
2502.04790 | Yuting Zeng | Yuting Zeng, Weizhe Huang, Lei Jiang, Tongxuan Liu, Xitai Jin, Chen
Tianying Tiana, Jing Li, Xiaohua Xu | S$^2$-MAD: Breaking the Token Barrier to Enhance Multi-Agent Debate
Efficiency | Accepted to NAACL 2025 Main | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have demonstrated remarkable capabilities across
various natural language processing (NLP) scenarios, but they still face
challenges when handling complex arithmetic and logical reasoning tasks. While
Chain-Of-Thought (CoT) reasoning, self-consistency (SC) and self-correction
strategies have attempted to guide models in sequential, multi-step reasoning,
Multi-agent Debate (MAD) has emerged as a viable approach for enhancing the
reasoning capabilities of LLMs. By increasing both the number of agents and the
frequency of debates, the performance of LLMs improves significantly. However,
this strategy results in a significant increase in token costs, presenting a
barrier to scalability. To address this challenge, we introduce a novel
sparsification strategy designed to reduce token costs within MAD. This
approach minimizes ineffective exchanges of information and unproductive
discussions among agents, thereby enhancing the overall efficiency of the
debate process. We conduct comparative experiments on multiple datasets across
various models, demonstrating that our approach significantly reduces the token
costs in MAD to a considerable extent. Specifically, compared to MAD, our
approach achieves an impressive reduction of up to 94.5\% in token costs while
maintaining performance degradation below 2.0\%.
| [
{
"version": "v1",
"created": "Fri, 7 Feb 2025 09:49:56 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 02:29:35 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Zeng",
"Yuting",
""
],
[
"Huang",
"Weizhe",
""
],
[
"Jiang",
"Lei",
""
],
[
"Liu",
"Tongxuan",
""
],
[
"Jin",
"Xitai",
""
],
[
"Tiana",
"Chen Tianying",
""
],
[
"Li",
"Jing",
""
],
[
"Xu",
"Xiaohua",
""
]
] | TITLE: S$^2$-MAD: Breaking the Token Barrier to Enhance Multi-Agent Debate
Efficiency
ABSTRACT: Large language models (LLMs) have demonstrated remarkable capabilities across
various natural language processing (NLP) scenarios, but they still face
challenges when handling complex arithmetic and logical reasoning tasks. While
Chain-Of-Thought (CoT) reasoning, self-consistency (SC) and self-correction
strategies have attempted to guide models in sequential, multi-step reasoning,
Multi-agent Debate (MAD) has emerged as a viable approach for enhancing the
reasoning capabilities of LLMs. By increasing both the number of agents and the
frequency of debates, the performance of LLMs improves significantly. However,
this strategy results in a significant increase in token costs, presenting a
barrier to scalability. To address this challenge, we introduce a novel
sparsification strategy designed to reduce token costs within MAD. This
approach minimizes ineffective exchanges of information and unproductive
discussions among agents, thereby enhancing the overall efficiency of the
debate process. We conduct comparative experiments on multiple datasets across
various models, demonstrating that our approach significantly reduces the token
costs in MAD to a considerable extent. Specifically, compared to MAD, our
approach achieves an impressive reduction of up to 94.5\% in token costs while
maintaining performance degradation below 2.0\%.
|
2502.05780 | Danny Wang | Danny Wang, Ruihong Qiu, Guangdong Bai, Zi Huang | GOLD: Graph Out-of-Distribution Detection via Implicit Adversarial
Latent Generation | ICLR25 | null | null | null | cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Despite graph neural networks' (GNNs) great success in modelling
graph-structured data, out-of-distribution (OOD) test instances still pose a
great challenge for current GNNs. One of the most effective techniques to
detect OOD nodes is to expose the detector model with an additional OOD
node-set, yet the extra OOD instances are often difficult to obtain in
practice. Recent methods for image data address this problem using OOD data
synthesis, typically relying on pre-trained generative models like Stable
Diffusion. However, these approaches require vast amounts of additional data,
as well as one-for-all pre-trained generative models, which are not available
for graph data. Therefore, we propose the GOLD framework for graph OOD
detection, an implicit adversarial learning pipeline with synthetic OOD
exposure without pre-trained models. The implicit adversarial training process
employs a novel alternating optimisation framework by training: (1) a latent
generative model to regularly imitate the in-distribution (ID) embeddings from
an evolving GNN, and (2) a GNN encoder and an OOD detector to accurately
classify ID data while increasing the energy divergence between the ID
embeddings and the generative model's synthetic embeddings. This novel approach
implicitly transforms the synthetic embeddings into pseudo-OOD instances
relative to the ID data, effectively simulating exposure to OOD scenarios
without auxiliary data. Extensive OOD detection experiments are conducted on
five benchmark graph datasets, verifying the superior performance of GOLD
without using real OOD data compared with the state-of-the-art OOD exposure and
non-exposure baselines.
| [
{
"version": "v1",
"created": "Sun, 9 Feb 2025 05:19:53 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 11:41:23 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Wang",
"Danny",
""
],
[
"Qiu",
"Ruihong",
""
],
[
"Bai",
"Guangdong",
""
],
[
"Huang",
"Zi",
""
]
] | TITLE: GOLD: Graph Out-of-Distribution Detection via Implicit Adversarial
Latent Generation
ABSTRACT: Despite graph neural networks' (GNNs) great success in modelling
graph-structured data, out-of-distribution (OOD) test instances still pose a
great challenge for current GNNs. One of the most effective techniques to
detect OOD nodes is to expose the detector model with an additional OOD
node-set, yet the extra OOD instances are often difficult to obtain in
practice. Recent methods for image data address this problem using OOD data
synthesis, typically relying on pre-trained generative models like Stable
Diffusion. However, these approaches require vast amounts of additional data,
as well as one-for-all pre-trained generative models, which are not available
for graph data. Therefore, we propose the GOLD framework for graph OOD
detection, an implicit adversarial learning pipeline with synthetic OOD
exposure without pre-trained models. The implicit adversarial training process
employs a novel alternating optimisation framework by training: (1) a latent
generative model to regularly imitate the in-distribution (ID) embeddings from
an evolving GNN, and (2) a GNN encoder and an OOD detector to accurately
classify ID data while increasing the energy divergence between the ID
embeddings and the generative model's synthetic embeddings. This novel approach
implicitly transforms the synthetic embeddings into pseudo-OOD instances
relative to the ID data, effectively simulating exposure to OOD scenarios
without auxiliary data. Extensive OOD detection experiments are conducted on
five benchmark graph datasets, verifying the superior performance of GOLD
without using real OOD data compared with the state-of-the-art OOD exposure and
non-exposure baselines.
|
2502.06193 | Ruiqi Wang | Ruiqi Wang, Jiyu Guo, Cuiyun Gao, Guodong Fan, Chun Yong Chong, Xin
Xia | Can LLMs Replace Human Evaluators? An Empirical Study of LLM-as-a-Judge
in Software Engineering | Accepted by ISSTA 2025:
https://conf.researchr.org/details/issta-2025/issta-2025-papers/85/Can-LLMs-replace-Human-Evaluators-An-Empirical-Study-of-LLM-as-a-Judge-in-Software-E | null | 10.1145/3728963 | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recently, large language models (LLMs) have been deployed to tackle various
software engineering (SE) tasks like code generation, significantly advancing
the automation of SE tasks. However, assessing the quality of these
LLM-generated code and text remains challenging. The commonly used Pass@k
metric necessitates extensive unit tests and configured environments, demands a
high labor cost, and is not suitable for evaluating LLM-generated text.
Conventional metrics like BLEU, which measure only lexical rather than semantic
similarity, have also come under scrutiny. In response, a new trend has emerged
to employ LLMs for automated evaluation, known as LLM-as-a-judge. These
LLM-as-a-judge methods are claimed to better mimic human assessment than
conventional metrics without relying on high-quality reference answers.
Nevertheless, their exact human alignment in SE tasks remains unexplored. In
this paper, we empirically explore LLM-as-a-judge methods for evaluating SE
tasks, focusing on their alignment with human judgments. We select seven
LLM-as-a-judge methods that utilize general-purpose LLMs, alongside two LLMs
specifically fine-tuned for evaluation. After generating and manually scoring
LLM responses on three recent SE datasets of code translation, code generation,
and code summarization, we then prompt these methods to evaluate each response.
Finally, we compare the scores generated by these methods with human
evaluation. The results indicate that output-based methods reach the highest
Pearson correlation of 81.32 and 68.51 with human scores in code translation
and generation, achieving near-human evaluation, noticeably outperforming
ChrF++, one of the best conventional metrics, at 34.23 and 64.92. Such
output-based methods prompt LLMs to output judgments directly, and exhibit more
balanced score distributions that resemble human score patterns. Finally, we
provide...
| [
{
"version": "v1",
"created": "Mon, 10 Feb 2025 06:49:29 GMT"
},
{
"version": "v2",
"created": "Thu, 10 Apr 2025 07:33:55 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Wang",
"Ruiqi",
""
],
[
"Guo",
"Jiyu",
""
],
[
"Gao",
"Cuiyun",
""
],
[
"Fan",
"Guodong",
""
],
[
"Chong",
"Chun Yong",
""
],
[
"Xia",
"Xin",
""
]
] | TITLE: Can LLMs Replace Human Evaluators? An Empirical Study of LLM-as-a-Judge
in Software Engineering
ABSTRACT: Recently, large language models (LLMs) have been deployed to tackle various
software engineering (SE) tasks like code generation, significantly advancing
the automation of SE tasks. However, assessing the quality of these
LLM-generated code and text remains challenging. The commonly used Pass@k
metric necessitates extensive unit tests and configured environments, demands a
high labor cost, and is not suitable for evaluating LLM-generated text.
Conventional metrics like BLEU, which measure only lexical rather than semantic
similarity, have also come under scrutiny. In response, a new trend has emerged
to employ LLMs for automated evaluation, known as LLM-as-a-judge. These
LLM-as-a-judge methods are claimed to better mimic human assessment than
conventional metrics without relying on high-quality reference answers.
Nevertheless, their exact human alignment in SE tasks remains unexplored. In
this paper, we empirically explore LLM-as-a-judge methods for evaluating SE
tasks, focusing on their alignment with human judgments. We select seven
LLM-as-a-judge methods that utilize general-purpose LLMs, alongside two LLMs
specifically fine-tuned for evaluation. After generating and manually scoring
LLM responses on three recent SE datasets of code translation, code generation,
and code summarization, we then prompt these methods to evaluate each response.
Finally, we compare the scores generated by these methods with human
evaluation. The results indicate that output-based methods reach the highest
Pearson correlation of 81.32 and 68.51 with human scores in code translation
and generation, achieving near-human evaluation, noticeably outperforming
ChrF++, one of the best conventional metrics, at 34.23 and 64.92. Such
output-based methods prompt LLMs to output judgments directly, and exhibit more
balanced score distributions that resemble human score patterns. Finally, we
provide...
|
2502.07532 | Erik Larsson | Erik Larsson, Joel Oskarsson, Tomas Landelius, Fredrik Lindsten | Diffusion-LAM: Probabilistic Limited Area Weather Forecasting with
Diffusion | Accepted, camera ready version | null | null | null | cs.LG physics.ao-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning methods have been shown to be effective for weather
forecasting, based on the speed and accuracy compared to traditional numerical
models. While early efforts primarily concentrated on deterministic
predictions, the field has increasingly shifted toward probabilistic
forecasting to better capture the forecast uncertainty. Most machine
learning-based models have been designed for global-scale predictions, with
only limited work targeting regional or limited area forecasting, which allows
more specialized and flexible modeling for specific locations. This work
introduces Diffusion-LAM, a probabilistic limited area weather model leveraging
conditional diffusion. By conditioning on boundary data from surrounding
regions, our approach generates forecasts within a defined area. Experimental
results on the MEPS limited area dataset demonstrate the potential of
Diffusion-LAM to deliver accurate probabilistic forecasts, highlighting its
promise for limited-area weather prediction.
| [
{
"version": "v1",
"created": "Tue, 11 Feb 2025 13:15:16 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Feb 2025 13:55:08 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Apr 2025 12:10:33 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Larsson",
"Erik",
""
],
[
"Oskarsson",
"Joel",
""
],
[
"Landelius",
"Tomas",
""
],
[
"Lindsten",
"Fredrik",
""
]
] | TITLE: Diffusion-LAM: Probabilistic Limited Area Weather Forecasting with
Diffusion
ABSTRACT: Machine learning methods have been shown to be effective for weather
forecasting, based on the speed and accuracy compared to traditional numerical
models. While early efforts primarily concentrated on deterministic
predictions, the field has increasingly shifted toward probabilistic
forecasting to better capture the forecast uncertainty. Most machine
learning-based models have been designed for global-scale predictions, with
only limited work targeting regional or limited area forecasting, which allows
more specialized and flexible modeling for specific locations. This work
introduces Diffusion-LAM, a probabilistic limited area weather model leveraging
conditional diffusion. By conditioning on boundary data from surrounding
regions, our approach generates forecasts within a defined area. Experimental
results on the MEPS limited area dataset demonstrate the potential of
Diffusion-LAM to deliver accurate probabilistic forecasts, highlighting its
promise for limited-area weather prediction.
|
2502.09535 | Carlton Shepherd | Carlton Shepherd and Elliot Hurley | Entropy Collapse in Mobile Sensors: The Hidden Risks of Sensor-Based
Security | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | Mobile sensor data has been proposed for security-critical applications such
as device pairing, proximity detection, and continuous authentication. However,
the foundational premise that these signals provide sufficient entropy remains
under-explored. In this work, we systematically analyse the entropy of mobile
sensor data across four diverse datasets spanning multiple application
contexts. Our findings reveal pervasive biases, with single-sensor mean
min-entropy values ranging from 3.408-4.483 bits (S.D.=1.018-1.574) despite
Shannon entropy being several multiples higher, showing a significant collapse
between average- to worst-case settings. We further demonstrate that
correlations between sensor modalities reduce the worst-case entropy of using
multiple sensors by up to ~75% compared to average-case Shannon entropy. This
brings joint min-entropy well below 10 bits in many cases and, in the best
case, yielding only ~24 bits of min-entropy when combining 20 sensor
modalities. These results raise the serious risk of attacks that exhaustively
search the space of possible sensor measurements. Our work also calls into
question the widely held assumption that adding more sensors inherently yields
higher security, and we strongly urge caution when relying on mobile sensor
data for security applications.
| [
{
"version": "v1",
"created": "Thu, 13 Feb 2025 17:50:58 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Feb 2025 10:43:47 GMT"
},
{
"version": "v3",
"created": "Mon, 17 Feb 2025 22:41:20 GMT"
},
{
"version": "v4",
"created": "Tue, 25 Mar 2025 11:42:52 GMT"
},
{
"version": "v5",
"created": "Wed, 26 Mar 2025 22:14:50 GMT"
},
{
"version": "v6",
"created": "Thu, 10 Apr 2025 17:53:17 GMT"
}
] | 2025-04-11T00:00:00 | [
[
"Shepherd",
"Carlton",
""
],
[
"Hurley",
"Elliot",
""
]
] | TITLE: Entropy Collapse in Mobile Sensors: The Hidden Risks of Sensor-Based
Security
ABSTRACT: Mobile sensor data has been proposed for security-critical applications such
as device pairing, proximity detection, and continuous authentication. However,
the foundational premise that these signals provide sufficient entropy remains
under-explored. In this work, we systematically analyse the entropy of mobile
sensor data across four diverse datasets spanning multiple application
contexts. Our findings reveal pervasive biases, with single-sensor mean
min-entropy values ranging from 3.408-4.483 bits (S.D.=1.018-1.574) despite
Shannon entropy being several multiples higher, showing a significant collapse
between average- to worst-case settings. We further demonstrate that
correlations between sensor modalities reduce the worst-case entropy of using
multiple sensors by up to ~75% compared to average-case Shannon entropy. This
brings joint min-entropy well below 10 bits in many cases and, in the best
case, yielding only ~24 bits of min-entropy when combining 20 sensor
modalities. These results raise the serious risk of attacks that exhaustively
search the space of possible sensor measurements. Our work also calls into
question the widely held assumption that adding more sensors inherently yields
higher security, and we strongly urge caution when relying on mobile sensor
data for security applications.
|
End of preview. Expand
in Data Studio
README.md exists but content is empty.
- Downloads last month
- 92