id
stringlengths 9
16
| submitter
stringlengths 3
64
⌀ | authors
stringlengths 5
6.63k
| title
stringlengths 7
245
| comments
stringlengths 1
482
⌀ | journal-ref
stringlengths 4
382
⌀ | doi
stringlengths 9
151
⌀ | report-no
stringclasses 984
values | categories
stringlengths 5
108
| license
stringclasses 9
values | abstract
stringlengths 83
3.41k
| versions
listlengths 1
20
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-04-11 00:00:00
| authors_parsed
listlengths 1
427
| prompt
stringlengths 166
3.49k
| label
stringclasses 2
values | prob
float64 0.5
0.98
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2503.00666
|
Ki-Hwan Oh
|
Ki-Hwan Oh, Leonardo Borgioli, Milo\v{s} \v{Z}efran, Valentina Valle,
Pier Cristoforo Giulianotti
|
Autonomous Dissection in Robotic Cholecystectomy
|
Submitted for IEEE/RSJ International Conference on Intelligent Robots
and Systems (IROS) 2025
| null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Robotic surgery offers enhanced precision and adaptability, paving the way
for automation in surgical interventions. Cholecystectomy, the gallbladder
removal, is particularly well-suited for automation due to its standardized
procedural steps and distinct anatomical boundaries. A key challenge in
automating this procedure is dissecting with accuracy and adaptability. This
paper presents a vision-based autonomous robotic dissection architecture that
integrates real-time segmentation, keypoint detection, grasping and stretching
the gallbladder with the left arm, and dissecting with the other. We introduce
an improved segmentation dataset based on videos of robotic cholecystectomy
performed by various surgeons, incorporating a new ``liver bed'' class to
enhance boundary tracking after multiple rounds of dissection. Our system
employs state-of-the-art segmentation models and an adaptive boundary
extraction method that maintains accuracy despite tissue deformations and
visual variations. Moreover, we implemented an automated grasping and pulling
strategy to optimize tissue tension before dissection upon our previous work.
Ex vivo evaluations on porcine livers demonstrate that our framework
significantly improves dissection precision and consistency, marking a step
toward fully autonomous robotic cholecystectomy.
|
[
{
"version": "v1",
"created": "Sat, 1 Mar 2025 23:38:19 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Oh",
"Ki-Hwan",
""
],
[
"Borgioli",
"Leonardo",
""
],
[
"Žefran",
"Miloš",
""
],
[
"Valle",
"Valentina",
""
],
[
"Giulianotti",
"Pier Cristoforo",
""
]
] |
TITLE: Autonomous Dissection in Robotic Cholecystectomy
ABSTRACT: Robotic surgery offers enhanced precision and adaptability, paving the way
for automation in surgical interventions. Cholecystectomy, the gallbladder
removal, is particularly well-suited for automation due to its standardized
procedural steps and distinct anatomical boundaries. A key challenge in
automating this procedure is dissecting with accuracy and adaptability. This
paper presents a vision-based autonomous robotic dissection architecture that
integrates real-time segmentation, keypoint detection, grasping and stretching
the gallbladder with the left arm, and dissecting with the other. We introduce
an improved segmentation dataset based on videos of robotic cholecystectomy
performed by various surgeons, incorporating a new ``liver bed'' class to
enhance boundary tracking after multiple rounds of dissection. Our system
employs state-of-the-art segmentation models and an adaptive boundary
extraction method that maintains accuracy despite tissue deformations and
visual variations. Moreover, we implemented an automated grasping and pulling
strategy to optimize tissue tension before dissection upon our previous work.
Ex vivo evaluations on porcine livers demonstrate that our framework
significantly improves dissection precision and consistency, marking a step
toward fully autonomous robotic cholecystectomy.
|
new_dataset
| 0.954605 |
2503.00670
|
Debashis Sen
|
Gargi V. Pillai, Ashish Verma and Debashis Sen
|
Transformer Based Self-Context Aware Prediction for Few-Shot Anomaly
Detection in Videos
|
Copyright 2022 IEEE. Personal use of this material is permitted.
Permission from IEEE must be obtained for all other uses, in any current or
future media, including reprinting/republishing this material for advertising
or promotional purposes, creating new collective works, for resale or
redistribution to servers or lists, or reuse of any copyrighted component of
this work in other works
|
IEEE International Conference on Image Processing (ICIP),
Bordeaux, France, 2022, pp. 3485-3489
|
10.1109/ICIP46576.2022.9897615
| null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Anomaly detection in videos is a challenging task as anomalies in different
videos are of different kinds. Therefore, a promising way to approach video
anomaly detection is by learning the non-anomalous nature of the video at hand.
To this end, we propose a one-class few-shot learning driven transformer based
approach for anomaly detection in videos that is self-context aware. Features
from the first few consecutive non-anomalous frames in a video are used to
train the transformer in predicting the non-anomalous feature of the subsequent
frame. This takes place under the attention of a self-context learned from the
input features themselves. After the learning, given a few previous frames, the
video-specific transformer is used to infer if a frame is anomalous or not by
comparing the feature predicted by it with the actual. The effectiveness of the
proposed method with respect to the state-of-the-art is demonstrated through
qualitative and quantitative results on different standard datasets. We also
study the positive effect of the self-context used in our approach.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 00:07:49 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Pillai",
"Gargi V.",
""
],
[
"Verma",
"Ashish",
""
],
[
"Sen",
"Debashis",
""
]
] |
TITLE: Transformer Based Self-Context Aware Prediction for Few-Shot Anomaly
Detection in Videos
ABSTRACT: Anomaly detection in videos is a challenging task as anomalies in different
videos are of different kinds. Therefore, a promising way to approach video
anomaly detection is by learning the non-anomalous nature of the video at hand.
To this end, we propose a one-class few-shot learning driven transformer based
approach for anomaly detection in videos that is self-context aware. Features
from the first few consecutive non-anomalous frames in a video are used to
train the transformer in predicting the non-anomalous feature of the subsequent
frame. This takes place under the attention of a self-context learned from the
input features themselves. After the learning, given a few previous frames, the
video-specific transformer is used to infer if a frame is anomalous or not by
comparing the feature predicted by it with the actual. The effectiveness of the
proposed method with respect to the state-of-the-art is demonstrated through
qualitative and quantitative results on different standard datasets. We also
study the positive effect of the self-context used in our approach.
|
no_new_dataset
| 0.950134 |
2503.00673
|
Pouya Fathollahzadeh
|
Pouya Fathollahzadeh, Mariam El Mezouar, Hao Li, Ying Zou, Ahmed E.
Hassan
|
Towards Refining Developer Questions using LLM-Based Named Entity
Recognition for Developer Chatroom Conversations
| null | null | null | null |
cs.SE cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
In software engineering chatrooms, communication is often hindered by
imprecise questions that cannot be answered. Recognizing key entities can be
essential for improving question clarity and facilitating better exchange.
However, existing research using natural language processing techniques often
overlooks these software-specific nuances. In this paper, we introduce
Software-specific Named Entity Recognition, Intent Detection, and Resolution
Classification (SENIR), a labeling approach that leverages a Large Language
Model to annotate entities, intents, and resolution status in developer
chatroom conversations. To offer quantitative guidance for improving question
clarity and resolvability, we build a resolution prediction model that
leverages SENIR's entity and intent labels along with additional predictive
features. We evaluate SENIR on the DISCO dataset using a subset of annotated
chatroom dialogues. SENIR achieves an 86% F-score for entity recognition, a 71%
F-score for intent detection, and an 89% F-score for resolution status
classification. Furthermore, our resolution prediction model, tested with
various sampling strategies (random undersampling and oversampling with SMOTE)
and evaluation methods (5-fold cross-validation, 10-fold cross-validation, and
bootstrapping), demonstrates AUC values ranging from 0.7 to 0.8. Key factors
influencing resolution include positive sentiment and entities such as
Programming Language and User Variable across multiple intents, while
diagnostic entities are more relevant in error-related questions. Moreover,
resolution rates vary significantly by intent: questions about API Usage and
API Change achieve higher resolution rates, whereas Discrepancy and Review have
lower resolution rates. A Chi-Square analysis confirms the statistical
significance of these differences.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 00:20:24 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Fathollahzadeh",
"Pouya",
""
],
[
"Mezouar",
"Mariam El",
""
],
[
"Li",
"Hao",
""
],
[
"Zou",
"Ying",
""
],
[
"Hassan",
"Ahmed E.",
""
]
] |
TITLE: Towards Refining Developer Questions using LLM-Based Named Entity
Recognition for Developer Chatroom Conversations
ABSTRACT: In software engineering chatrooms, communication is often hindered by
imprecise questions that cannot be answered. Recognizing key entities can be
essential for improving question clarity and facilitating better exchange.
However, existing research using natural language processing techniques often
overlooks these software-specific nuances. In this paper, we introduce
Software-specific Named Entity Recognition, Intent Detection, and Resolution
Classification (SENIR), a labeling approach that leverages a Large Language
Model to annotate entities, intents, and resolution status in developer
chatroom conversations. To offer quantitative guidance for improving question
clarity and resolvability, we build a resolution prediction model that
leverages SENIR's entity and intent labels along with additional predictive
features. We evaluate SENIR on the DISCO dataset using a subset of annotated
chatroom dialogues. SENIR achieves an 86% F-score for entity recognition, a 71%
F-score for intent detection, and an 89% F-score for resolution status
classification. Furthermore, our resolution prediction model, tested with
various sampling strategies (random undersampling and oversampling with SMOTE)
and evaluation methods (5-fold cross-validation, 10-fold cross-validation, and
bootstrapping), demonstrates AUC values ranging from 0.7 to 0.8. Key factors
influencing resolution include positive sentiment and entities such as
Programming Language and User Variable across multiple intents, while
diagnostic entities are more relevant in error-related questions. Moreover,
resolution rates vary significantly by intent: questions about API Usage and
API Change achieve higher resolution rates, whereas Discrepancy and Review have
lower resolution rates. A Chi-Square analysis confirms the statistical
significance of these differences.
|
no_new_dataset
| 0.957278 |
2503.00674
|
Yan Wang
|
Yan Wang, Lingfei Qian, Xueqing Peng, Jimin Huang, Dongji Feng
|
OrdRankBen: A Novel Ranking Benchmark for Ordinal Relevance in NLP
|
6 pages
| null | null | null |
cs.IR cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The evaluation of ranking tasks remains a significant challenge in natural
language processing (NLP), particularly due to the lack of direct labels for
results in real-world scenarios. Benchmark datasets play a crucial role in
providing standardized testbeds that ensure fair comparisons, enhance
reproducibility, and enable progress tracking, facilitating rigorous assessment
and continuous improvement of ranking models. Existing NLP ranking benchmarks
typically use binary relevance labels or continuous relevance scores,
neglecting ordinal relevance scores. However, binary labels oversimplify
relevance distinctions, while continuous scores lack a clear ordinal structure,
making it challenging to capture nuanced ranking differences effectively. To
address these challenges, we introduce OrdRankBen, a novel benchmark designed
to capture multi-granularity relevance distinctions. Unlike conventional
benchmarks, OrdRankBen incorporates structured ordinal labels, enabling more
precise ranking evaluations. Given the absence of suitable datasets for ordinal
relevance ranking in NLP, we constructed two datasets with distinct ordinal
label distributions. We further evaluate various models for three model types,
ranking-based language models, general large language models, and
ranking-focused large language models on these datasets. Experimental results
show that ordinal relevance modeling provides a more precise evaluation of
ranking models, improving their ability to distinguish multi-granularity
differences among ranked items-crucial for tasks that demand fine-grained
relevance differentiation.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 00:28:55 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Wang",
"Yan",
""
],
[
"Qian",
"Lingfei",
""
],
[
"Peng",
"Xueqing",
""
],
[
"Huang",
"Jimin",
""
],
[
"Feng",
"Dongji",
""
]
] |
TITLE: OrdRankBen: A Novel Ranking Benchmark for Ordinal Relevance in NLP
ABSTRACT: The evaluation of ranking tasks remains a significant challenge in natural
language processing (NLP), particularly due to the lack of direct labels for
results in real-world scenarios. Benchmark datasets play a crucial role in
providing standardized testbeds that ensure fair comparisons, enhance
reproducibility, and enable progress tracking, facilitating rigorous assessment
and continuous improvement of ranking models. Existing NLP ranking benchmarks
typically use binary relevance labels or continuous relevance scores,
neglecting ordinal relevance scores. However, binary labels oversimplify
relevance distinctions, while continuous scores lack a clear ordinal structure,
making it challenging to capture nuanced ranking differences effectively. To
address these challenges, we introduce OrdRankBen, a novel benchmark designed
to capture multi-granularity relevance distinctions. Unlike conventional
benchmarks, OrdRankBen incorporates structured ordinal labels, enabling more
precise ranking evaluations. Given the absence of suitable datasets for ordinal
relevance ranking in NLP, we constructed two datasets with distinct ordinal
label distributions. We further evaluate various models for three model types,
ranking-based language models, general large language models, and
ranking-focused large language models on these datasets. Experimental results
show that ordinal relevance modeling provides a more precise evaluation of
ranking models, improving their ability to distinguish multi-granularity
differences among ranked items-crucial for tasks that demand fine-grained
relevance differentiation.
|
new_dataset
| 0.97066 |
2503.00676
|
Rishikesh Joshi
|
Rishikesh Joshi and Junaed Sattar
|
One-Shot Gesture Recognition for Underwater Diver-To-Robot Communication
|
17 pages, 8 figures, 2 tables, submitted to IROS2025
| null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Reliable human-robot communication is essential for underwater human-robot
interaction (U-HRI), yet traditional methods such as acoustic signaling and
predefined gesture-based models suffer from limitations in adaptability and
robustness. In this work, we propose One-Shot Gesture Recognition (OSG), a
novel method that enables real-time, pose-based, temporal gesture recognition
underwater from a single demonstration, eliminating the need for extensive
dataset collection or model retraining. OSG leverages shape-based
classification techniques, including Hu moments, Zernike moments, and Fourier
descriptors, to robustly recognize gestures in visually-challenging underwater
environments. Our system achieves high accuracy on real-world underwater data
and operates efficiently on embedded hardware commonly found on autonomous
underwater vehicles (AUVs), demonstrating its feasibility for deployment
on-board robots. Compared to deep learning approaches, OSG is lightweight,
computationally efficient, and highly adaptable, making it ideal for
diver-to-robot communication. We evaluate OSG's performance on an augmented
gesture dataset and real-world underwater video data, comparing its accuracy
against deep learning methods. Our results show OSG's potential to enhance
U-HRI by enabling the immediate deployment of user-defined gestures without the
constraints of predefined gesture languages.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 00:52:55 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Joshi",
"Rishikesh",
""
],
[
"Sattar",
"Junaed",
""
]
] |
TITLE: One-Shot Gesture Recognition for Underwater Diver-To-Robot Communication
ABSTRACT: Reliable human-robot communication is essential for underwater human-robot
interaction (U-HRI), yet traditional methods such as acoustic signaling and
predefined gesture-based models suffer from limitations in adaptability and
robustness. In this work, we propose One-Shot Gesture Recognition (OSG), a
novel method that enables real-time, pose-based, temporal gesture recognition
underwater from a single demonstration, eliminating the need for extensive
dataset collection or model retraining. OSG leverages shape-based
classification techniques, including Hu moments, Zernike moments, and Fourier
descriptors, to robustly recognize gestures in visually-challenging underwater
environments. Our system achieves high accuracy on real-world underwater data
and operates efficiently on embedded hardware commonly found on autonomous
underwater vehicles (AUVs), demonstrating its feasibility for deployment
on-board robots. Compared to deep learning approaches, OSG is lightweight,
computationally efficient, and highly adaptable, making it ideal for
diver-to-robot communication. We evaluate OSG's performance on an augmented
gesture dataset and real-world underwater video data, comparing its accuracy
against deep learning methods. Our results show OSG's potential to enhance
U-HRI by enabling the immediate deployment of user-defined gestures without the
constraints of predefined gesture languages.
|
no_new_dataset
| 0.947527 |
2503.00686
|
Leming Shen
|
Leming Shen, Qiang Yang, Xinyu Huang, Zijing Ma, Yuanqing Zheng
|
GPIoT: Tailoring Small Language Models for IoT Program Synthesis and
Development
| null | null | null | null |
cs.SE cs.AI
|
http://creativecommons.org/publicdomain/zero/1.0/
|
Code Large Language Models (LLMs) enhance software development efficiency by
automatically generating code and documentation in response to user
requirements. However, code LLMs cannot synthesize specialized programs when
tasked with IoT applications that require domain knowledge. While
Retrieval-Augmented Generation (RAG) offers a promising solution by fetching
relevant domain knowledge, it necessitates powerful cloud LLMs (e.g., GPT-4) to
process user requirements and retrieved contents, which raises significant
privacy concerns. This approach also suffers from unstable networks and
prohibitive LLM query costs. Moreover, it is challenging to ensure the
correctness and relevance of the fetched contents. To address these issues, we
propose GPIoT, a code generation system for IoT applications by fine-tuning
locally deployable Small Language Models (SLMs) on IoT-specialized datasets.
SLMs have smaller model sizes, allowing efficient local deployment and
execution to mitigate privacy concerns and network uncertainty. Furthermore, by
fine-tuning the SLMs with our IoT-specialized datasets, the SLMs' ability to
synthesize IoT-related programs can be substantially improved. To evaluate
GPIoT's capability in synthesizing programs for IoT applications, we develop a
benchmark, IoTBench. Extensive experiments and user trials demonstrate the
effectiveness of GPIoT in generating IoT-specialized code, outperforming
state-of-the-art code LLMs with an average task accuracy increment of 64.7% and
significant improvements in user satisfaction.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 01:55:40 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Shen",
"Leming",
""
],
[
"Yang",
"Qiang",
""
],
[
"Huang",
"Xinyu",
""
],
[
"Ma",
"Zijing",
""
],
[
"Zheng",
"Yuanqing",
""
]
] |
TITLE: GPIoT: Tailoring Small Language Models for IoT Program Synthesis and
Development
ABSTRACT: Code Large Language Models (LLMs) enhance software development efficiency by
automatically generating code and documentation in response to user
requirements. However, code LLMs cannot synthesize specialized programs when
tasked with IoT applications that require domain knowledge. While
Retrieval-Augmented Generation (RAG) offers a promising solution by fetching
relevant domain knowledge, it necessitates powerful cloud LLMs (e.g., GPT-4) to
process user requirements and retrieved contents, which raises significant
privacy concerns. This approach also suffers from unstable networks and
prohibitive LLM query costs. Moreover, it is challenging to ensure the
correctness and relevance of the fetched contents. To address these issues, we
propose GPIoT, a code generation system for IoT applications by fine-tuning
locally deployable Small Language Models (SLMs) on IoT-specialized datasets.
SLMs have smaller model sizes, allowing efficient local deployment and
execution to mitigate privacy concerns and network uncertainty. Furthermore, by
fine-tuning the SLMs with our IoT-specialized datasets, the SLMs' ability to
synthesize IoT-related programs can be substantially improved. To evaluate
GPIoT's capability in synthesizing programs for IoT applications, we develop a
benchmark, IoTBench. Extensive experiments and user trials demonstrate the
effectiveness of GPIoT in generating IoT-specialized code, outperforming
state-of-the-art code LLMs with an average task accuracy increment of 64.7% and
significant improvements in user satisfaction.
|
no_new_dataset
| 0.917043 |
2503.00689
|
Amir Mohammad Mirzaei
|
Amir Mohammad Mirzaei
|
Stress, Strain, or Displacement? A Novel Machine Learning Based
Framework to Predict Mixed Mode I/II Fracture Toughness
| null | null | null | null |
physics.comp-ph cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Accurate prediction of fracture toughness under complex loading conditions,
like mixed mode I/II, is essential for reliable failure assessment. This paper
aims to develop a machine learning framework for predicting fracture toughness
and crack initiation angles by directly utilizing stress, strain, or
displacement distributions represented by selected nodes as input features.
Validation is conducted using experimental data across various mode mixities
and specimen geometries for brittle materials. Among stress, strain, and
displacement fields, it is shown that the stress-based features, when paired
with Multilayer Perceptron models, achieve high predictive accuracy with R2
scores exceeding 0.86 for fracture load predictions and 0.94 for angle
predictions. A comparison with the Theory of Critical Distances (Generalized
Maximum Tangential Stress) demonstrates the high accuracy of the framework.
Furthermore, the impact of input parameter selections is studied, and it is
demonstrated that advanced feature selection algorithms enable the framework to
handle different ranges and densities of the representing field. The
framework's performance was further validated for datasets with a limited
number of data points and restricted mode mixities, where it maintained high
accuracy. The proposed framework is computationally efficient and practical,
and it operates without any supplementary post-processing steps, such as stress
intensity factor calculations.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 02:02:11 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Mirzaei",
"Amir Mohammad",
""
]
] |
TITLE: Stress, Strain, or Displacement? A Novel Machine Learning Based
Framework to Predict Mixed Mode I/II Fracture Toughness
ABSTRACT: Accurate prediction of fracture toughness under complex loading conditions,
like mixed mode I/II, is essential for reliable failure assessment. This paper
aims to develop a machine learning framework for predicting fracture toughness
and crack initiation angles by directly utilizing stress, strain, or
displacement distributions represented by selected nodes as input features.
Validation is conducted using experimental data across various mode mixities
and specimen geometries for brittle materials. Among stress, strain, and
displacement fields, it is shown that the stress-based features, when paired
with Multilayer Perceptron models, achieve high predictive accuracy with R2
scores exceeding 0.86 for fracture load predictions and 0.94 for angle
predictions. A comparison with the Theory of Critical Distances (Generalized
Maximum Tangential Stress) demonstrates the high accuracy of the framework.
Furthermore, the impact of input parameter selections is studied, and it is
demonstrated that advanced feature selection algorithms enable the framework to
handle different ranges and densities of the representing field. The
framework's performance was further validated for datasets with a limited
number of data points and restricted mode mixities, where it maintained high
accuracy. The proposed framework is computationally efficient and practical,
and it operates without any supplementary post-processing steps, such as stress
intensity factor calculations.
|
no_new_dataset
| 0.946498 |
2503.00697
|
Yiyang Lin
|
Yiyang Lin, Danling Jiang, Xinyu Liu, Yun Miao, and Yixuan Yuan
|
CREATE-FFPE: Cross-Resolution Compensated and Multi-Frequency Enhanced
FS-to-FFPE Stain Transfer for Intraoperative IHC Images
| null | null | null | null |
cs.CV cs.AI eess.IV
|
http://creativecommons.org/licenses/by/4.0/
|
In the immunohistochemical (IHC) analysis during surgery, frozen-section (FS)
images are used to determine the benignity or malignancy of the tumor. However,
FS image faces problems such as image contamination and poor nuclear detail,
which may disturb the pathologist's diagnosis. In contrast, formalin-fixed and
paraffin-embedded (FFPE) image has a higher staining quality, but it requires
quite a long time to prepare and thus is not feasible during surgery. To help
pathologists observe IHC images with high quality in surgery, this paper
proposes a Cross-REsolution compensATed and multi-frequency Enhanced FS-to-FFPE
(CREATE-FFPE) stain transfer framework, which is the first FS-to-FFPE method
for the intraoperative IHC images. To solve the slide contamination and poor
nuclear detail mentioned above, we propose the cross-resolution compensation
module (CRCM) and the wavelet detail guidance module (WDGM). Specifically, CRCM
compensates for information loss due to contamination by providing more tissue
information across multiple resolutions, while WDGM produces the desirable
details in a wavelet way, and the details can be used to guide the stain
transfer to be more precise. Experiments show our method can beat all the
competing methods on our dataset. In addition, the FID has decreased by 44.4%,
and KID*100 has decreased by 71.2% by adding the proposed CRCM and WDGM in
ablation studies, and the performance of a downstream microsatellite
instability prediction task with public dataset can be greatly improved by
performing our FS-to-FFPE stain transfer.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 02:38:11 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Lin",
"Yiyang",
""
],
[
"Jiang",
"Danling",
""
],
[
"Liu",
"Xinyu",
""
],
[
"Miao",
"Yun",
""
],
[
"Yuan",
"Yixuan",
""
]
] |
TITLE: CREATE-FFPE: Cross-Resolution Compensated and Multi-Frequency Enhanced
FS-to-FFPE Stain Transfer for Intraoperative IHC Images
ABSTRACT: In the immunohistochemical (IHC) analysis during surgery, frozen-section (FS)
images are used to determine the benignity or malignancy of the tumor. However,
FS image faces problems such as image contamination and poor nuclear detail,
which may disturb the pathologist's diagnosis. In contrast, formalin-fixed and
paraffin-embedded (FFPE) image has a higher staining quality, but it requires
quite a long time to prepare and thus is not feasible during surgery. To help
pathologists observe IHC images with high quality in surgery, this paper
proposes a Cross-REsolution compensATed and multi-frequency Enhanced FS-to-FFPE
(CREATE-FFPE) stain transfer framework, which is the first FS-to-FFPE method
for the intraoperative IHC images. To solve the slide contamination and poor
nuclear detail mentioned above, we propose the cross-resolution compensation
module (CRCM) and the wavelet detail guidance module (WDGM). Specifically, CRCM
compensates for information loss due to contamination by providing more tissue
information across multiple resolutions, while WDGM produces the desirable
details in a wavelet way, and the details can be used to guide the stain
transfer to be more precise. Experiments show our method can beat all the
competing methods on our dataset. In addition, the FID has decreased by 44.4%,
and KID*100 has decreased by 71.2% by adding the proposed CRCM and WDGM in
ablation studies, and the performance of a downstream microsatellite
instability prediction task with public dataset can be greatly improved by
performing our FS-to-FFPE stain transfer.
|
no_new_dataset
| 0.944638 |
2503.00711
|
Zhijiang Wan
|
Zhijiang Wan, Qianhao Yu, Jia Mao, Wenfeng Duan and Cheng Ding
|
OpenECG: Benchmarking ECG Foundation Models with Public 1.2 Million
Records
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
This study introduces OpenECG, a large-scale benchmark of 1.2 million 12-lead
ECG recordings from nine centers, to evaluate ECG foundation models (ECG-FMs)
trained on public datasets. We investigate three self-supervised learning
methods (SimCLR, BYOL, MAE) with ResNet-50 and Vision Transformer
architectures, assessing model generalization through leave-one-dataset-out
experiments and data scaling analysis. Results show that pre-training on
diverse datasets significantly improves generalization, with BYOL and MAE
outperforming SimCLR, highlighting the efficacy of feature-consistency and
generative learning over contrastive approaches. Data scaling experiments
reveal that performance saturates at 60-70% of total data for BYOL and MAE,
while SimCLR requires more data. These findings demonstrate that publicly
available ECG data can match or surpass proprietary datasets in training robust
ECG-FMs, paving the way for scalable, clinically meaningful AI-driven ECG
analysis.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 03:26:14 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Wan",
"Zhijiang",
""
],
[
"Yu",
"Qianhao",
""
],
[
"Mao",
"Jia",
""
],
[
"Duan",
"Wenfeng",
""
],
[
"Ding",
"Cheng",
""
]
] |
TITLE: OpenECG: Benchmarking ECG Foundation Models with Public 1.2 Million
Records
ABSTRACT: This study introduces OpenECG, a large-scale benchmark of 1.2 million 12-lead
ECG recordings from nine centers, to evaluate ECG foundation models (ECG-FMs)
trained on public datasets. We investigate three self-supervised learning
methods (SimCLR, BYOL, MAE) with ResNet-50 and Vision Transformer
architectures, assessing model generalization through leave-one-dataset-out
experiments and data scaling analysis. Results show that pre-training on
diverse datasets significantly improves generalization, with BYOL and MAE
outperforming SimCLR, highlighting the efficacy of feature-consistency and
generative learning over contrastive approaches. Data scaling experiments
reveal that performance saturates at 60-70% of total data for BYOL and MAE,
while SimCLR requires more data. These findings demonstrate that publicly
available ECG data can match or surpass proprietary datasets in training robust
ECG-FMs, paving the way for scalable, clinically meaningful AI-driven ECG
analysis.
|
no_new_dataset
| 0.944022 |
2503.00714
|
Haoyu Li
|
Haoyu Li, Srikanth Kandula, Maria Angels de Luis Balaguer, Aditya
Akella, Venkat Arun
|
Speculative Ad-hoc Querying
| null | null | null | null |
cs.DB cs.AI cs.HC cs.LG cs.MA
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Analyzing large datasets requires responsive query execution, but executing
SQL queries on massive datasets can be slow. This paper explores whether query
execution can begin even before the user has finished typing, allowing results
to appear almost instantly. We propose SpeQL, a system that leverages Large
Language Models (LLMs) to predict likely queries based on the database schema,
the user's past queries, and their incomplete query. Since exact query
prediction is infeasible, SpeQL speculates on partial queries in two ways: 1)
it predicts the query structure to compile and plan queries in advance, and 2)
it precomputes smaller temporary tables that are much smaller than the original
database, but are still predicted to contain all information necessary to
answer the user's final query. Additionally, SpeQL continuously displays
results for speculated queries and subqueries in real time, aiding exploratory
analysis. A utility/user study showed that SpeQL improved task completion time,
and participants reported that its speculative display of results helped them
discover patterns in the data more quickly. In the study, SpeQL improves user's
query latency by up to $289\times$ and kept the overhead reasonable, at $\$4$
per hour.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 03:44:31 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Li",
"Haoyu",
""
],
[
"Kandula",
"Srikanth",
""
],
[
"Balaguer",
"Maria Angels de Luis",
""
],
[
"Akella",
"Aditya",
""
],
[
"Arun",
"Venkat",
""
]
] |
TITLE: Speculative Ad-hoc Querying
ABSTRACT: Analyzing large datasets requires responsive query execution, but executing
SQL queries on massive datasets can be slow. This paper explores whether query
execution can begin even before the user has finished typing, allowing results
to appear almost instantly. We propose SpeQL, a system that leverages Large
Language Models (LLMs) to predict likely queries based on the database schema,
the user's past queries, and their incomplete query. Since exact query
prediction is infeasible, SpeQL speculates on partial queries in two ways: 1)
it predicts the query structure to compile and plan queries in advance, and 2)
it precomputes smaller temporary tables that are much smaller than the original
database, but are still predicted to contain all information necessary to
answer the user's final query. Additionally, SpeQL continuously displays
results for speculated queries and subqueries in real time, aiding exploratory
analysis. A utility/user study showed that SpeQL improved task completion time,
and participants reported that its speculative display of results helped them
discover patterns in the data more quickly. In the study, SpeQL improves user's
query latency by up to $289\times$ and kept the overhead reasonable, at $\$4$
per hour.
|
no_new_dataset
| 0.941654 |
2503.00731
|
Yang Ding
|
Yang Ding, Can Han, Sijia Du, Yaqi Wang, Dahong Qian
|
LightEndoStereo: A Real-time Lightweight Stereo Matching Method for
Endoscopy Images
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Real-time acquisition of accurate depth of scene is essential for automated
robotic minimally invasive surgery, and stereo matching with binocular
endoscopy can generate such depth. However, existing algorithms struggle with
ambiguous tissue boundaries and real-time performance in prevalent
high-resolution endoscopic scenes. We propose LightEndoStereo, a lightweight
real-time stereo matching method for endoscopic images. We introduce a 3D Mamba
Coordinate Attention module to streamline the cost aggregation process by
generating position-sensitive attention maps and capturing long-range
dependencies across spatial dimensions using the Mamba block. Additionally, we
introduce a High-Frequency Disparity Optimization module to refine disparity
estimates at tissue boundaries by enhancing high-frequency information in the
wavelet domain. Our method is evaluated on the SCARED and SERV-CT datasets,
achieving state-of-the-art matching accuracy and a real-time inference speed of
42 FPS. The code is available at https://github.com/Sonne-Ding/LightEndoStereo.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 05:06:52 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ding",
"Yang",
""
],
[
"Han",
"Can",
""
],
[
"Du",
"Sijia",
""
],
[
"Wang",
"Yaqi",
""
],
[
"Qian",
"Dahong",
""
]
] |
TITLE: LightEndoStereo: A Real-time Lightweight Stereo Matching Method for
Endoscopy Images
ABSTRACT: Real-time acquisition of accurate depth of scene is essential for automated
robotic minimally invasive surgery, and stereo matching with binocular
endoscopy can generate such depth. However, existing algorithms struggle with
ambiguous tissue boundaries and real-time performance in prevalent
high-resolution endoscopic scenes. We propose LightEndoStereo, a lightweight
real-time stereo matching method for endoscopic images. We introduce a 3D Mamba
Coordinate Attention module to streamline the cost aggregation process by
generating position-sensitive attention maps and capturing long-range
dependencies across spatial dimensions using the Mamba block. Additionally, we
introduce a High-Frequency Disparity Optimization module to refine disparity
estimates at tissue boundaries by enhancing high-frequency information in the
wavelet domain. Our method is evaluated on the SCARED and SERV-CT datasets,
achieving state-of-the-art matching accuracy and a real-time inference speed of
42 FPS. The code is available at https://github.com/Sonne-Ding/LightEndoStereo.
|
no_new_dataset
| 0.949949 |
2503.00737
|
Jinjiang You
|
Jinjiang You, Hewei Wang, Yijie Li, Mingxiao Huo, Long Van Tran Ha,
Mingyuan Ma, Jinfeng Xu, Puzhen Wu, Shubham Garg, Wei Pu
|
Multi-Cali Anything: Dense Feature Multi-Frame Structure-from-Motion for
Large-Scale Camera Array Calibration
|
8 pages
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Calibrating large-scale camera arrays, such as those in dome-based setups, is
time-intensive and typically requires dedicated captures of known patterns.
While extrinsics in such arrays are fixed due to the physical setup, intrinsics
often vary across sessions due to factors like lens adjustments or temperature
changes. In this paper, we propose a dense-feature-driven multi-frame
calibration method that refines intrinsics directly from scene data,
eliminating the necessity for additional calibration captures. Our approach
enhances traditional Structure-from-Motion (SfM) pipelines by introducing an
extrinsics regularization term to progressively align estimated extrinsics with
ground-truth values, a dense feature reprojection term to reduce keypoint
errors by minimizing reprojection loss in the feature space, and an intrinsics
variance term for joint optimization across multiple frames. Experiments on the
Multiface dataset show that our method achieves nearly the same precision as
dedicated calibration processes, and significantly enhances intrinsics and 3D
reconstruction accuracy. Fully compatible with existing SfM pipelines, our
method provides an efficient and practical plug-and-play solution for
large-scale camera setups. Our code is publicly available at:
https://github.com/YJJfish/Multi-Cali-Anything
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 05:25:17 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"You",
"Jinjiang",
""
],
[
"Wang",
"Hewei",
""
],
[
"Li",
"Yijie",
""
],
[
"Huo",
"Mingxiao",
""
],
[
"Ha",
"Long Van Tran",
""
],
[
"Ma",
"Mingyuan",
""
],
[
"Xu",
"Jinfeng",
""
],
[
"Wu",
"Puzhen",
""
],
[
"Garg",
"Shubham",
""
],
[
"Pu",
"Wei",
""
]
] |
TITLE: Multi-Cali Anything: Dense Feature Multi-Frame Structure-from-Motion for
Large-Scale Camera Array Calibration
ABSTRACT: Calibrating large-scale camera arrays, such as those in dome-based setups, is
time-intensive and typically requires dedicated captures of known patterns.
While extrinsics in such arrays are fixed due to the physical setup, intrinsics
often vary across sessions due to factors like lens adjustments or temperature
changes. In this paper, we propose a dense-feature-driven multi-frame
calibration method that refines intrinsics directly from scene data,
eliminating the necessity for additional calibration captures. Our approach
enhances traditional Structure-from-Motion (SfM) pipelines by introducing an
extrinsics regularization term to progressively align estimated extrinsics with
ground-truth values, a dense feature reprojection term to reduce keypoint
errors by minimizing reprojection loss in the feature space, and an intrinsics
variance term for joint optimization across multiple frames. Experiments on the
Multiface dataset show that our method achieves nearly the same precision as
dedicated calibration processes, and significantly enhances intrinsics and 3D
reconstruction accuracy. Fully compatible with existing SfM pipelines, our
method provides an efficient and practical plug-and-play solution for
large-scale camera setups. Our code is publicly available at:
https://github.com/YJJfish/Multi-Cali-Anything
|
no_new_dataset
| 0.949435 |
2503.00744
|
Anyang Ji
|
Anyang Ji, Qingbo Kang, Wei Xu, Changfan Wang, Kang Li and Qicheng Lao
|
Confounder-Aware Medical Data Selection for Fine-Tuning Pretrained
Vision Models
|
5 pages, 3 figures
| null | null | null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The emergence of large-scale pre-trained vision foundation models has greatly
advanced the medical imaging field through the pre-training and fine-tuning
paradigm. However, selecting appropriate medical data for downstream
fine-tuning remains a significant challenge considering its annotation cost,
privacy concerns, and the detrimental effects of confounding variables. In this
work, we present a confounder-aware medical data selection approach for medical
dataset curation aiming to select minimal representative data by strategically
mitigating the undesirable impact of confounding variables while preserving the
natural distribution of the dataset. Our approach first identifies confounding
variables within data and then develops a distance-based data selection
strategy for confounder-aware sampling with a constrained budget in the data
size. We validate the superiority of our approach through extensive experiments
across diverse medical imaging modalities, highlighting its effectiveness in
addressing the substantial impact of confounding variables and enhancing the
fine-tuning efficiency in the medical imaging domain, compared to other data
selection approaches.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 05:50:25 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ji",
"Anyang",
""
],
[
"Kang",
"Qingbo",
""
],
[
"Xu",
"Wei",
""
],
[
"Wang",
"Changfan",
""
],
[
"Li",
"Kang",
""
],
[
"Lao",
"Qicheng",
""
]
] |
TITLE: Confounder-Aware Medical Data Selection for Fine-Tuning Pretrained
Vision Models
ABSTRACT: The emergence of large-scale pre-trained vision foundation models has greatly
advanced the medical imaging field through the pre-training and fine-tuning
paradigm. However, selecting appropriate medical data for downstream
fine-tuning remains a significant challenge considering its annotation cost,
privacy concerns, and the detrimental effects of confounding variables. In this
work, we present a confounder-aware medical data selection approach for medical
dataset curation aiming to select minimal representative data by strategically
mitigating the undesirable impact of confounding variables while preserving the
natural distribution of the dataset. Our approach first identifies confounding
variables within data and then develops a distance-based data selection
strategy for confounder-aware sampling with a constrained budget in the data
size. We validate the superiority of our approach through extensive experiments
across diverse medical imaging modalities, highlighting its effectiveness in
addressing the substantial impact of confounding variables and enhancing the
fine-tuning efficiency in the medical imaging domain, compared to other data
selection approaches.
|
no_new_dataset
| 0.953622 |
2503.00748
|
Xiangde Luo
|
Zihao Luo, Zijun Gao, Wenjun Liao, Shichuan Zhang, Guotai Wang, and
Xiangde Luo
|
Dynamic Gradient Sparsification Training for Few-Shot Fine-tuning of CT
Lymph Node Segmentation Foundation Model
|
10 pages, 3 figures, 2 tables, and the lymph node segmentation
foundation model code and pretrained model are available
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Accurate lymph node (LN) segmentation is critical in radiotherapy treatment
and prognosis analysis, but is limited by the need for large annotated
datasets. While deep learning-based segmentation foundation models show
potential in developing high-performing models with fewer samples, their
medical adaptation faces LN domain-specific prior deficiencies and inefficient
few-shot fine-tuning for complex clinical practices, highlighting the necessity
of an LN segmentation foundation model. In this work, we annotated 36,106
visible LNs from 3,346 publicly available head-and-neck CT scans to establish a
robust LN segmentation model (nnUNetv2). Building on this, we propose Dynamic
Gradient Sparsification Training (DGST), a few-shot fine-tuning approach that
preserves foundational knowledge while dynamically updating the most critical
parameters of the LN segmentation model with few annotations. We validate it on
two publicly available LN segmentation datasets: SegRap2023 and LNQ2023. The
results show that DGST outperforms existing few-shot fine-tuning methods,
achieving satisfactory performance with limited labeled data. We release the
dataset, models and all implementations to facilitate relevant research:
https://github.com/Zihaoluoh/LN-Seg-FM.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 06:02:34 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Luo",
"Zihao",
""
],
[
"Gao",
"Zijun",
""
],
[
"Liao",
"Wenjun",
""
],
[
"Zhang",
"Shichuan",
""
],
[
"Wang",
"Guotai",
""
],
[
"Luo",
"Xiangde",
""
]
] |
TITLE: Dynamic Gradient Sparsification Training for Few-Shot Fine-tuning of CT
Lymph Node Segmentation Foundation Model
ABSTRACT: Accurate lymph node (LN) segmentation is critical in radiotherapy treatment
and prognosis analysis, but is limited by the need for large annotated
datasets. While deep learning-based segmentation foundation models show
potential in developing high-performing models with fewer samples, their
medical adaptation faces LN domain-specific prior deficiencies and inefficient
few-shot fine-tuning for complex clinical practices, highlighting the necessity
of an LN segmentation foundation model. In this work, we annotated 36,106
visible LNs from 3,346 publicly available head-and-neck CT scans to establish a
robust LN segmentation model (nnUNetv2). Building on this, we propose Dynamic
Gradient Sparsification Training (DGST), a few-shot fine-tuning approach that
preserves foundational knowledge while dynamically updating the most critical
parameters of the LN segmentation model with few annotations. We validate it on
two publicly available LN segmentation datasets: SegRap2023 and LNQ2023. The
results show that DGST outperforms existing few-shot fine-tuning methods,
achieving satisfactory performance with limited labeled data. We release the
dataset, models and all implementations to facilitate relevant research:
https://github.com/Zihaoluoh/LN-Seg-FM.
|
no_new_dataset
| 0.506454 |
2503.00750
|
Xingbo Fu
|
Xingbo Fu, Yinhan He, Jundong Li
|
Edge Prompt Tuning for Graph Neural Networks
|
Accepted by ICLR 2025
| null | null | null |
cs.LG cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Pre-training powerful Graph Neural Networks (GNNs) with unlabeled graph data
in a self-supervised manner has emerged as a prominent technique in recent
years. However, inevitable objective gaps often exist between pre-training and
downstream tasks. To bridge this gap, graph prompt tuning techniques design and
learn graph prompts by manipulating input graphs or reframing downstream tasks
as pre-training tasks without fine-tuning the pre-trained GNN models. While
recent graph prompt tuning methods have proven effective in adapting
pre-trained GNN models for downstream tasks, they overlook the crucial role of
edges in graph prompt design, which can significantly affect the quality of
graph representations for downstream tasks. In this study, we propose
EdgePrompt, a simple yet effective graph prompt tuning method from the
perspective of edges. Unlike previous studies that design prompt vectors on
node features, EdgePrompt manipulates input graphs by learning additional
prompt vectors for edges and incorporates the edge prompts through message
passing in the pre-trained GNN models to better embed graph structural
information for downstream tasks. Our method is compatible with prevalent GNN
architectures pre-trained under various pre-training strategies and is
universal for different downstream tasks. We provide comprehensive theoretical
analyses of our method regarding its capability of handling node classification
and graph classification as downstream tasks. Extensive experiments on ten
graph datasets under four pre-training strategies demonstrate the superiority
of our proposed method against six baselines. Our code is available at
https://github.com/xbfu/EdgePrompt.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 06:07:54 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Fu",
"Xingbo",
""
],
[
"He",
"Yinhan",
""
],
[
"Li",
"Jundong",
""
]
] |
TITLE: Edge Prompt Tuning for Graph Neural Networks
ABSTRACT: Pre-training powerful Graph Neural Networks (GNNs) with unlabeled graph data
in a self-supervised manner has emerged as a prominent technique in recent
years. However, inevitable objective gaps often exist between pre-training and
downstream tasks. To bridge this gap, graph prompt tuning techniques design and
learn graph prompts by manipulating input graphs or reframing downstream tasks
as pre-training tasks without fine-tuning the pre-trained GNN models. While
recent graph prompt tuning methods have proven effective in adapting
pre-trained GNN models for downstream tasks, they overlook the crucial role of
edges in graph prompt design, which can significantly affect the quality of
graph representations for downstream tasks. In this study, we propose
EdgePrompt, a simple yet effective graph prompt tuning method from the
perspective of edges. Unlike previous studies that design prompt vectors on
node features, EdgePrompt manipulates input graphs by learning additional
prompt vectors for edges and incorporates the edge prompts through message
passing in the pre-trained GNN models to better embed graph structural
information for downstream tasks. Our method is compatible with prevalent GNN
architectures pre-trained under various pre-training strategies and is
universal for different downstream tasks. We provide comprehensive theoretical
analyses of our method regarding its capability of handling node classification
and graph classification as downstream tasks. Extensive experiments on ten
graph datasets under four pre-training strategies demonstrate the superiority
of our proposed method against six baselines. Our code is available at
https://github.com/xbfu/EdgePrompt.
|
no_new_dataset
| 0.94801 |
2503.00751
|
Hongchao Gu
|
Hongchao Gu, Dexun Li, Kuicai Dong, Hao Zhang, Hang Lv, Hao Wang, Defu
Lian, Yong Liu, Enhong Chen
|
RAPID: Efficient Retrieval-Augmented Long Text Generation with Writing
Planning and Information Discovery
| null | null | null | null |
cs.CL cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Generating knowledge-intensive and comprehensive long texts, such as
encyclopedia articles, remains significant challenges for Large Language
Models. It requires not only the precise integration of facts but also the
maintenance of thematic coherence throughout the article. Existing methods,
such as direct generation and multi-agent discussion, often struggle with
issues like hallucinations, topic incoherence, and significant latency. To
address these challenges, we propose RAPID, an efficient retrieval-augmented
long text generation framework. RAPID consists of three main modules: (1)
Retrieval-augmented preliminary outline generation to reduce hallucinations,
(2) Attribute-constrained search for efficient information discovery, (3)
Plan-guided article generation for enhanced coherence. Extensive experiments on
our newly compiled benchmark dataset, FreshWiki-2024, demonstrate that RAPID
significantly outperforms state-of-the-art methods across a wide range of
evaluation metrics (e.g. long-text generation, outline quality, latency, etc).
Our work provides a robust and efficient solution to the challenges of
automated long-text generation.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 06:11:29 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Gu",
"Hongchao",
""
],
[
"Li",
"Dexun",
""
],
[
"Dong",
"Kuicai",
""
],
[
"Zhang",
"Hao",
""
],
[
"Lv",
"Hang",
""
],
[
"Wang",
"Hao",
""
],
[
"Lian",
"Defu",
""
],
[
"Liu",
"Yong",
""
],
[
"Chen",
"Enhong",
""
]
] |
TITLE: RAPID: Efficient Retrieval-Augmented Long Text Generation with Writing
Planning and Information Discovery
ABSTRACT: Generating knowledge-intensive and comprehensive long texts, such as
encyclopedia articles, remains significant challenges for Large Language
Models. It requires not only the precise integration of facts but also the
maintenance of thematic coherence throughout the article. Existing methods,
such as direct generation and multi-agent discussion, often struggle with
issues like hallucinations, topic incoherence, and significant latency. To
address these challenges, we propose RAPID, an efficient retrieval-augmented
long text generation framework. RAPID consists of three main modules: (1)
Retrieval-augmented preliminary outline generation to reduce hallucinations,
(2) Attribute-constrained search for efficient information discovery, (3)
Plan-guided article generation for enhanced coherence. Extensive experiments on
our newly compiled benchmark dataset, FreshWiki-2024, demonstrate that RAPID
significantly outperforms state-of-the-art methods across a wide range of
evaluation metrics (e.g. long-text generation, outline quality, latency, etc).
Our work provides a robust and efficient solution to the challenges of
automated long-text generation.
|
new_dataset
| 0.959345 |
2503.00760
|
Lei Zhou
|
Lei Zhou, Nimu Yuan, Katjana Ehrlich, Jinyi Qi
|
NCF: Neural Correspondence Field for Medical Image Registration
| null | null | null | null |
eess.IV cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Deformable image registration is a fundamental task in medical image
processing. Traditional optimization-based methods often struggle with accuracy
in dealing with complex deformation. Recently, learning-based methods have
achieved good performance on public datasets, but the scarcity of medical image
data makes it challenging to build a generalizable model to handle diverse
real-world scenarios. To address this, we propose a training-data-free
learning-based method, Neural Correspondence Field (NCF), which can learn from
just one data pair. Our approach employs a compact neural network to model the
correspondence field and optimize model parameters for each individual image
pair. Consequently, each pair has a unique set of network weights. Notably, our
model is highly efficient, utilizing only 0.06 million parameters. Evaluation
results showed that the proposed method achieved superior performance on a
public Lung CT dataset and outperformed a traditional method on a head and neck
dataset, demonstrating both its effectiveness and efficiency.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 06:55:49 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhou",
"Lei",
""
],
[
"Yuan",
"Nimu",
""
],
[
"Ehrlich",
"Katjana",
""
],
[
"Qi",
"Jinyi",
""
]
] |
TITLE: NCF: Neural Correspondence Field for Medical Image Registration
ABSTRACT: Deformable image registration is a fundamental task in medical image
processing. Traditional optimization-based methods often struggle with accuracy
in dealing with complex deformation. Recently, learning-based methods have
achieved good performance on public datasets, but the scarcity of medical image
data makes it challenging to build a generalizable model to handle diverse
real-world scenarios. To address this, we propose a training-data-free
learning-based method, Neural Correspondence Field (NCF), which can learn from
just one data pair. Our approach employs a compact neural network to model the
correspondence field and optimize model parameters for each individual image
pair. Consequently, each pair has a unique set of network weights. Notably, our
model is highly efficient, utilizing only 0.06 million parameters. Evaluation
results showed that the proposed method achieved superior performance on a
public Lung CT dataset and outperformed a traditional method on a head and neck
dataset, demonstrating both its effectiveness and efficiency.
|
no_new_dataset
| 0.952574 |
2503.00771
|
Yupu Hao
|
Yupu Hao, Pengfei Cao, Zhuoran Jin, Huanxuan Liao, Yubo Chen, Kang
Liu, Jun Zhao
|
Evaluating Personalized Tool-Augmented LLMs from the Perspectives of
Personalization and Proactivity
| null | null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Personalized tool utilization is essential for aligning large language models
(LLMs) with user preference in interaction scenarios with various tools.
However, most of the current benchmarks primarily focus on either
personalization of text generation or direct tool-utilizing, without
considering both. In this work, we introduce a novel benchmark ETAPP for
evaluating personalized tool invocation, establishing a sandbox environment,
and a comprehensive dataset of 800 testing cases covering diverse user
profiles. To improve the accuracy of our evaluation, we propose a
key-point-based LLM evaluation method, mitigating biases in the LLM-as-a-judge
system by manually annotating key points for each test case and providing them
to LLM as the reference. Additionally, we evaluate the excellent LLMs and
provide an in-depth analysis. Furthermore, we investigate the impact of
different tool-invoking strategies on LLMs' personalization performance and the
effects of fine-tuning in our task. The effectiveness of our preference-setting
and key-point-based evaluation method is also validated. Our findings offer
insights into improving personalized LLM agents. Our Code is available at
https://github.com/hypasd-art/ETAPP.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 07:36:22 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Hao",
"Yupu",
""
],
[
"Cao",
"Pengfei",
""
],
[
"Jin",
"Zhuoran",
""
],
[
"Liao",
"Huanxuan",
""
],
[
"Chen",
"Yubo",
""
],
[
"Liu",
"Kang",
""
],
[
"Zhao",
"Jun",
""
]
] |
TITLE: Evaluating Personalized Tool-Augmented LLMs from the Perspectives of
Personalization and Proactivity
ABSTRACT: Personalized tool utilization is essential for aligning large language models
(LLMs) with user preference in interaction scenarios with various tools.
However, most of the current benchmarks primarily focus on either
personalization of text generation or direct tool-utilizing, without
considering both. In this work, we introduce a novel benchmark ETAPP for
evaluating personalized tool invocation, establishing a sandbox environment,
and a comprehensive dataset of 800 testing cases covering diverse user
profiles. To improve the accuracy of our evaluation, we propose a
key-point-based LLM evaluation method, mitigating biases in the LLM-as-a-judge
system by manually annotating key points for each test case and providing them
to LLM as the reference. Additionally, we evaluate the excellent LLMs and
provide an in-depth analysis. Furthermore, we investigate the impact of
different tool-invoking strategies on LLMs' personalization performance and the
effects of fine-tuning in our task. The effectiveness of our preference-setting
and key-point-based evaluation method is also validated. Our findings offer
insights into improving personalized LLM agents. Our Code is available at
https://github.com/hypasd-art/ETAPP.
|
new_dataset
| 0.970042 |
2503.00780
|
Astitva Kamble
|
Astitva Kamble, Vani Bandodkar, Saakshi Dharmadhikary, Veena Anand,
Pradyut Kumar Sanki, Mei X. Wu, Biswabandhu Jana
|
Enhanced Multi-Class Classification of Gastrointestinal Endoscopic
Images with Interpretable Deep Learning Model
| null | null | null | null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Endoscopy serves as an essential procedure for evaluating the
gastrointestinal (GI) tract and plays a pivotal role in identifying GI-related
disorders. Recent advancements in deep learning have demonstrated substantial
progress in detecting abnormalities through intricate models and data
augmentation methods.This research introduces a novel approach to enhance
classification accuracy using 8,000 labeled endoscopic images from the Kvasir
dataset, categorized into eight distinct classes. Leveraging EfficientNetB3 as
the backbone, the proposed architecture eliminates reliance on data
augmentation while preserving moderate model complexity. The model achieves a
test accuracy of 94.25%, alongside precision and recall of 94.29% and 94.24%
respectively. Furthermore, Local Interpretable Model-agnostic Explanation
(LIME) saliency maps are employed to enhance interpretability by defining
critical regions in the images that influenced model predictions. Overall, this
work highlights the importance of AI in advancing medical imaging by combining
high classification accuracy with interpretability.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 08:07:50 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Kamble",
"Astitva",
""
],
[
"Bandodkar",
"Vani",
""
],
[
"Dharmadhikary",
"Saakshi",
""
],
[
"Anand",
"Veena",
""
],
[
"Sanki",
"Pradyut Kumar",
""
],
[
"Wu",
"Mei X.",
""
],
[
"Jana",
"Biswabandhu",
""
]
] |
TITLE: Enhanced Multi-Class Classification of Gastrointestinal Endoscopic
Images with Interpretable Deep Learning Model
ABSTRACT: Endoscopy serves as an essential procedure for evaluating the
gastrointestinal (GI) tract and plays a pivotal role in identifying GI-related
disorders. Recent advancements in deep learning have demonstrated substantial
progress in detecting abnormalities through intricate models and data
augmentation methods.This research introduces a novel approach to enhance
classification accuracy using 8,000 labeled endoscopic images from the Kvasir
dataset, categorized into eight distinct classes. Leveraging EfficientNetB3 as
the backbone, the proposed architecture eliminates reliance on data
augmentation while preserving moderate model complexity. The model achieves a
test accuracy of 94.25%, alongside precision and recall of 94.29% and 94.24%
respectively. Furthermore, Local Interpretable Model-agnostic Explanation
(LIME) saliency maps are employed to enhance interpretability by defining
critical regions in the images that influenced model predictions. Overall, this
work highlights the importance of AI in advancing medical imaging by combining
high classification accuracy with interpretability.
|
no_new_dataset
| 0.946151 |
2503.00790
|
Donghyun Yoon
|
Juho Lee, Donghyun Yoon, Gumoon Jeong, Hyeoncheol Kim
|
Acoustic Anomaly Detection on UAM Propeller Defect with Acoustic dataset
for Crack of drone Propeller (ADCP)
|
25 pages
| null | null | null |
cs.SD cs.ET eess.AS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The imminent commercialization of UAM requires stable, AI-based maintenance
systems to ensure safety for both passengers and pedestrians. This paper
presents a methodology for non-destructively detecting cracks in UAM propellers
using drone propeller sound datasets. Normal operating sounds were recorded,
and abnormal sounds (categorized as ripped and broken) were differentiated by
varying the microphone-propeller angle and throttle power. Our novel approach
integrates FFT and STFT preprocessing techniques to capture both global
frequency patterns and local time-frequency variations, thereby enhancing
anomaly detection performance. The constructed Acoustic Dataset for Crack of
Drone Propeller (ADCP) demonstrates the potential for detecting propeller
cracks and lays the groundwork for future UAM maintenance applications.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 08:40:23 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Lee",
"Juho",
""
],
[
"Yoon",
"Donghyun",
""
],
[
"Jeong",
"Gumoon",
""
],
[
"Kim",
"Hyeoncheol",
""
]
] |
TITLE: Acoustic Anomaly Detection on UAM Propeller Defect with Acoustic dataset
for Crack of drone Propeller (ADCP)
ABSTRACT: The imminent commercialization of UAM requires stable, AI-based maintenance
systems to ensure safety for both passengers and pedestrians. This paper
presents a methodology for non-destructively detecting cracks in UAM propellers
using drone propeller sound datasets. Normal operating sounds were recorded,
and abnormal sounds (categorized as ripped and broken) were differentiated by
varying the microphone-propeller angle and throttle power. Our novel approach
integrates FFT and STFT preprocessing techniques to capture both global
frequency patterns and local time-frequency variations, thereby enhancing
anomaly detection performance. The constructed Acoustic Dataset for Crack of
Drone Propeller (ADCP) demonstrates the potential for detecting propeller
cracks and lays the groundwork for future UAM maintenance applications.
|
new_dataset
| 0.952397 |
2503.00794
|
Kailun Yang
|
Longbin Zhang, Tsung-Lin Wu, Ananda Sidarta, Xiaoyue Yan, Prayook
Jatesiktat, Kailun Yang, Wei Tech Ang
|
Detecting Heel Strike and toe off Events Using Kinematic Methods and
LSTM Models
| null | null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Accurate gait event detection is crucial for gait analysis, rehabilitation,
and assistive technology, particularly in exoskeleton control, where precise
identification of stance and swing phases is essential. This study evaluated
the performance of seven kinematics-based methods and a Long Short-Term Memory
(LSTM) model for detecting heel strike and toe-off events across 4363 gait
cycles from 588 able-bodied subjects. The results indicated that while the Zeni
et al. method achieved the highest accuracy among kinematics-based approaches,
other methods exhibited systematic biases or required dataset-specific tuning.
The LSTM model performed comparably to Zeni et al., providing a data-driven
alternative without systematic bias. These findings highlight the potential of
deep learning-based approaches for gait event detection while emphasizing the
need for further validation in clinical populations and across diverse gait
conditions. Future research will explore the generalizability of these methods
in pathological populations, such as individuals with post-stroke conditions
and knee osteoarthritis, as well as their robustness across varied gait
conditions and data collection settings to enhance their applicability in
rehabilitation and exoskeleton control.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 08:46:13 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhang",
"Longbin",
""
],
[
"Wu",
"Tsung-Lin",
""
],
[
"Sidarta",
"Ananda",
""
],
[
"Yan",
"Xiaoyue",
""
],
[
"Jatesiktat",
"Prayook",
""
],
[
"Yang",
"Kailun",
""
],
[
"Ang",
"Wei Tech",
""
]
] |
TITLE: Detecting Heel Strike and toe off Events Using Kinematic Methods and
LSTM Models
ABSTRACT: Accurate gait event detection is crucial for gait analysis, rehabilitation,
and assistive technology, particularly in exoskeleton control, where precise
identification of stance and swing phases is essential. This study evaluated
the performance of seven kinematics-based methods and a Long Short-Term Memory
(LSTM) model for detecting heel strike and toe-off events across 4363 gait
cycles from 588 able-bodied subjects. The results indicated that while the Zeni
et al. method achieved the highest accuracy among kinematics-based approaches,
other methods exhibited systematic biases or required dataset-specific tuning.
The LSTM model performed comparably to Zeni et al., providing a data-driven
alternative without systematic bias. These findings highlight the potential of
deep learning-based approaches for gait event detection while emphasizing the
need for further validation in clinical populations and across diverse gait
conditions. Future research will explore the generalizability of these methods
in pathological populations, such as individuals with post-stroke conditions
and knee osteoarthritis, as well as their robustness across varied gait
conditions and data collection settings to enhance their applicability in
rehabilitation and exoskeleton control.
|
no_new_dataset
| 0.942082 |
2503.00801
|
Zikuan Li
|
Zikuan Li, Honghua Chen, Yuecheng Wang, Sibo Wu, Mingqiang Wei, Jun
Wang
|
STAR-Edge: Structure-aware Local Spherical Curve Representation for
Thin-walled Edge Extraction from Unstructured Point Clouds
|
Accepted at CVPR 2025
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Extracting geometric edges from unstructured point clouds remains a
significant challenge, particularly in thin-walled structures that are commonly
found in everyday objects. Traditional geometric methods and recent
learning-based approaches frequently struggle with these structures, as both
rely heavily on sufficient contextual information from local point
neighborhoods. However, 3D measurement data of thin-walled structures often
lack the accurate, dense, and regular neighborhood sampling required for
reliable edge extraction, resulting in degraded performance.
In this work, we introduce STAR-Edge, a novel approach designed for detecting
and refining edge points in thin-walled structures. Our method leverages a
unique representation-the local spherical curve-to create structure-aware
neighborhoods that emphasize co-planar points while reducing interference from
close-by, non-co-planar surfaces. This representation is transformed into a
rotation-invariant descriptor, which, combined with a lightweight multi-layer
perceptron, enables robust edge point classification even in the presence of
noise and sparse or irregular sampling. Besides, we also use the local
spherical curve representation to estimate more precise normals and introduce
an optimization function to project initially identified edge points exactly on
the true edges. Experiments conducted on the ABC dataset and thin-walled
structure-specific datasets demonstrate that STAR-Edge outperforms existing
edge detection methods, showcasing better robustness under various challenging
conditions.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 08:51:13 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Li",
"Zikuan",
""
],
[
"Chen",
"Honghua",
""
],
[
"Wang",
"Yuecheng",
""
],
[
"Wu",
"Sibo",
""
],
[
"Wei",
"Mingqiang",
""
],
[
"Wang",
"Jun",
""
]
] |
TITLE: STAR-Edge: Structure-aware Local Spherical Curve Representation for
Thin-walled Edge Extraction from Unstructured Point Clouds
ABSTRACT: Extracting geometric edges from unstructured point clouds remains a
significant challenge, particularly in thin-walled structures that are commonly
found in everyday objects. Traditional geometric methods and recent
learning-based approaches frequently struggle with these structures, as both
rely heavily on sufficient contextual information from local point
neighborhoods. However, 3D measurement data of thin-walled structures often
lack the accurate, dense, and regular neighborhood sampling required for
reliable edge extraction, resulting in degraded performance.
In this work, we introduce STAR-Edge, a novel approach designed for detecting
and refining edge points in thin-walled structures. Our method leverages a
unique representation-the local spherical curve-to create structure-aware
neighborhoods that emphasize co-planar points while reducing interference from
close-by, non-co-planar surfaces. This representation is transformed into a
rotation-invariant descriptor, which, combined with a lightweight multi-layer
perceptron, enables robust edge point classification even in the presence of
noise and sparse or irregular sampling. Besides, we also use the local
spherical curve representation to estimate more precise normals and introduce
an optimization function to project initially identified edge points exactly on
the true edges. Experiments conducted on the ABC dataset and thin-walled
structure-specific datasets demonstrate that STAR-Edge outperforms existing
edge detection methods, showcasing better robustness under various challenging
conditions.
|
no_new_dataset
| 0.951142 |
2503.00802
|
Jia-Xuan Jiang
|
Jia-Xuan Jiang, Wenhui Lei, Yifeng Wu, Hongtao Wu, Furong Li, Yining
Xie, Xiaofan Zhang, Zhong Wang
|
MFM-DA: Instance-Aware Adaptor and Hierarchical Alignment for Efficient
Domain Adaptation in Medical Foundation Models
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Medical Foundation Models (MFMs), trained on large-scale datasets, have
demonstrated superior performance across various tasks. However, these models
still struggle with domain gaps in practical applications. Specifically, even
after fine-tuning on source-domain data, task-adapted foundation models often
perform poorly in the target domain. To address this challenge, we propose a
few-shot unsupervised domain adaptation (UDA) framework for MFMs, named MFM-DA,
which only leverages a limited number of unlabeled target-domain images. Our
approach begins by training a Denoising Diffusion Probabilistic Model (DDPM),
which is then adapted to the target domain using a proposed dynamic
instance-aware adaptor and a distribution direction loss, enabling the DDPM to
translate source-domain images into the target domain style. The adapted images
are subsequently processed through the MFM, where we introduce a designed
channel-spatial alignment Low-Rank Adaptation (LoRA) to ensure effective
feature alignment. Extensive experiments on optic cup and disc segmentation
tasks demonstrate that MFM-DA outperforms state-of-the-art methods. Our work
provides a practical solution to the domain gap issue in real-world MFM
deployment. Code will be available at here.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 08:54:33 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Jiang",
"Jia-Xuan",
""
],
[
"Lei",
"Wenhui",
""
],
[
"Wu",
"Yifeng",
""
],
[
"Wu",
"Hongtao",
""
],
[
"Li",
"Furong",
""
],
[
"Xie",
"Yining",
""
],
[
"Zhang",
"Xiaofan",
""
],
[
"Wang",
"Zhong",
""
]
] |
TITLE: MFM-DA: Instance-Aware Adaptor and Hierarchical Alignment for Efficient
Domain Adaptation in Medical Foundation Models
ABSTRACT: Medical Foundation Models (MFMs), trained on large-scale datasets, have
demonstrated superior performance across various tasks. However, these models
still struggle with domain gaps in practical applications. Specifically, even
after fine-tuning on source-domain data, task-adapted foundation models often
perform poorly in the target domain. To address this challenge, we propose a
few-shot unsupervised domain adaptation (UDA) framework for MFMs, named MFM-DA,
which only leverages a limited number of unlabeled target-domain images. Our
approach begins by training a Denoising Diffusion Probabilistic Model (DDPM),
which is then adapted to the target domain using a proposed dynamic
instance-aware adaptor and a distribution direction loss, enabling the DDPM to
translate source-domain images into the target domain style. The adapted images
are subsequently processed through the MFM, where we introduce a designed
channel-spatial alignment Low-Rank Adaptation (LoRA) to ensure effective
feature alignment. Extensive experiments on optic cup and disc segmentation
tasks demonstrate that MFM-DA outperforms state-of-the-art methods. Our work
provides a practical solution to the domain gap issue in real-world MFM
deployment. Code will be available at here.
|
no_new_dataset
| 0.945601 |
2503.00803
|
Qingwen Zhang
|
Qingwen Zhang, Ajinkya Khoche, Yi Yang, Li Ling, Sina Sharif Mansouri,
Olov Andersson, Patric Jensfelt
|
HiMo: High-Speed Objects Motion Compensation in Point Clouds
|
12 pages
| null | null | null |
cs.CV cs.RO
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
LiDAR point clouds often contain motion-induced distortions, degrading the
accuracy of object appearances in the captured data. In this paper, we first
characterize the underlying reasons for the point cloud distortion and show
that this is present in public datasets. We find that this distortion is more
pronounced in high-speed environments such as highways, as well as in
multi-LiDAR configurations, a common setup for heavy vehicles. Previous work
has dealt with point cloud distortion from the ego-motion but fails to consider
distortion from the motion of other objects. We therefore introduce a novel
undistortion pipeline, HiMo, that leverages scene flow estimation for object
motion compensation, correcting the depiction of dynamic objects. We further
propose an extension of a state-of-the-art self-supervised scene flow method.
Due to the lack of well-established motion distortion metrics in the
literature, we also propose two metrics for compensation performance
evaluation: compensation accuracy at a point level and shape similarity on
objects. To demonstrate the efficacy of our method, we conduct extensive
experiments on the Argoverse 2 dataset and a new real-world dataset. Our new
dataset is collected from heavy vehicles equipped with multi-LiDARs and on
highways as opposed to mostly urban settings in the existing datasets. The
source code, including all methods and the evaluation data, will be provided
upon publication. See https://kin-zhang.github.io/HiMo for more details.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 08:55:12 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhang",
"Qingwen",
""
],
[
"Khoche",
"Ajinkya",
""
],
[
"Yang",
"Yi",
""
],
[
"Ling",
"Li",
""
],
[
"Mansouri",
"Sina Sharif",
""
],
[
"Andersson",
"Olov",
""
],
[
"Jensfelt",
"Patric",
""
]
] |
TITLE: HiMo: High-Speed Objects Motion Compensation in Point Clouds
ABSTRACT: LiDAR point clouds often contain motion-induced distortions, degrading the
accuracy of object appearances in the captured data. In this paper, we first
characterize the underlying reasons for the point cloud distortion and show
that this is present in public datasets. We find that this distortion is more
pronounced in high-speed environments such as highways, as well as in
multi-LiDAR configurations, a common setup for heavy vehicles. Previous work
has dealt with point cloud distortion from the ego-motion but fails to consider
distortion from the motion of other objects. We therefore introduce a novel
undistortion pipeline, HiMo, that leverages scene flow estimation for object
motion compensation, correcting the depiction of dynamic objects. We further
propose an extension of a state-of-the-art self-supervised scene flow method.
Due to the lack of well-established motion distortion metrics in the
literature, we also propose two metrics for compensation performance
evaluation: compensation accuracy at a point level and shape similarity on
objects. To demonstrate the efficacy of our method, we conduct extensive
experiments on the Argoverse 2 dataset and a new real-world dataset. Our new
dataset is collected from heavy vehicles equipped with multi-LiDARs and on
highways as opposed to mostly urban settings in the existing datasets. The
source code, including all methods and the evaluation data, will be provided
upon publication. See https://kin-zhang.github.io/HiMo for more details.
|
new_dataset
| 0.967656 |
2503.00807
|
Yuezhi Yang
|
Yuezhi Yang, Haitao Yang, Kiyohiro Nakayama, Xiangru Huang, Leonidas
Guibas, Qixing Huang
|
GenAnalysis: Joint Shape Analysis by Learning Man-Made Shape Generators
with Deformation Regularizations
|
21 pages, 25 figures
| null | null | null |
cs.GR cs.CV
|
http://creativecommons.org/licenses/by-sa/4.0/
|
We present GenAnalysis, an implicit shape generation framework that allows
joint analysis of man-made shapes, including shape matching and joint shape
segmentation. The key idea is to enforce an as-affine-as-possible (AAAP)
deformation between synthetic shapes of the implicit generator that are close
to each other in the latent space, which we achieve by designing a
regularization loss. It allows us to understand the shape variation of each
shape in the context of neighboring shapes and also offers structure-preserving
interpolations between the input shapes. We show how to extract these shape
variations by recovering piecewise affine vector fields in the tangent space of
each shape. These vector fields provide single-shape segmentation cues. We then
derive shape correspondences by iteratively propagating AAAP deformations
across a sequence of intermediate shapes. These correspondences are then used
to aggregate single-shape segmentation cues into consistent segmentations. We
conduct experiments on the ShapeNet dataset to show superior performance in
shape matching and joint shape segmentation over previous methods.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 09:17:08 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Yang",
"Yuezhi",
""
],
[
"Yang",
"Haitao",
""
],
[
"Nakayama",
"Kiyohiro",
""
],
[
"Huang",
"Xiangru",
""
],
[
"Guibas",
"Leonidas",
""
],
[
"Huang",
"Qixing",
""
]
] |
TITLE: GenAnalysis: Joint Shape Analysis by Learning Man-Made Shape Generators
with Deformation Regularizations
ABSTRACT: We present GenAnalysis, an implicit shape generation framework that allows
joint analysis of man-made shapes, including shape matching and joint shape
segmentation. The key idea is to enforce an as-affine-as-possible (AAAP)
deformation between synthetic shapes of the implicit generator that are close
to each other in the latent space, which we achieve by designing a
regularization loss. It allows us to understand the shape variation of each
shape in the context of neighboring shapes and also offers structure-preserving
interpolations between the input shapes. We show how to extract these shape
variations by recovering piecewise affine vector fields in the tangent space of
each shape. These vector fields provide single-shape segmentation cues. We then
derive shape correspondences by iteratively propagating AAAP deformations
across a sequence of intermediate shapes. These correspondences are then used
to aggregate single-shape segmentation cues into consistent segmentations. We
conduct experiments on the ShapeNet dataset to show superior performance in
shape matching and joint shape segmentation over previous methods.
|
no_new_dataset
| 0.946051 |
2503.00811
|
Lu Ma
|
Lu Ma, Kaibo Cao, Hao Liang, Jiaxin Lin, Zhuang Li, Yuhong Liu, Jihong
Zhang, Wentao Zhang, and Bin Cui
|
Evaluating and Predicting Distorted Human Body Parts for Generated
Images
|
8 pages, 6 figures
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Recent advancements in text-to-image (T2I) models enable high-quality image
synthesis, yet generating anatomically accurate human figures remains
challenging. AI-generated images frequently exhibit distortions such as
proliferated limbs, missing fingers, deformed extremities, or fused body parts.
Existing evaluation metrics like Inception Score (IS) and Fr\'echet Inception
Distance (FID) lack the granularity to detect these distortions, while human
preference-based metrics focus on abstract quality assessments rather than
anatomical fidelity. To address this gap, we establish the first standards for
identifying human body distortions in AI-generated images and introduce
Distortion-5K, a comprehensive dataset comprising 4,700 annotated images of
normal and malformed human figures across diverse styles and distortion types.
Based on this dataset, we propose ViT-HD, a Vision Transformer-based model
tailored for detecting human body distortions in AI-generated images, which
outperforms state-of-the-art segmentation models and visual language models,
achieving an F1 score of 0.899 and IoU of 0.831 on distortion localization.
Additionally, we construct the Human Distortion Benchmark with 500
human-centric prompts to evaluate four popular T2I models using trained ViT-HD,
revealing that nearly 50\% of generated images contain distortions. This work
pioneers a systematic approach to evaluating anatomical accuracy in
AI-generated humans, offering tools to advance the fidelity of T2I models and
their real-world applicability. The Distortion-5K dataset, trained ViT-HD will
soon be released in our GitHub repository:
\href{https://github.com/TheRoadQaQ/Predicting-Distortion}{https://github.com/TheRoadQaQ/Predicting-Distortion}.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 09:34:44 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ma",
"Lu",
""
],
[
"Cao",
"Kaibo",
""
],
[
"Liang",
"Hao",
""
],
[
"Lin",
"Jiaxin",
""
],
[
"Li",
"Zhuang",
""
],
[
"Liu",
"Yuhong",
""
],
[
"Zhang",
"Jihong",
""
],
[
"Zhang",
"Wentao",
""
],
[
"Cui",
"Bin",
""
]
] |
TITLE: Evaluating and Predicting Distorted Human Body Parts for Generated
Images
ABSTRACT: Recent advancements in text-to-image (T2I) models enable high-quality image
synthesis, yet generating anatomically accurate human figures remains
challenging. AI-generated images frequently exhibit distortions such as
proliferated limbs, missing fingers, deformed extremities, or fused body parts.
Existing evaluation metrics like Inception Score (IS) and Fr\'echet Inception
Distance (FID) lack the granularity to detect these distortions, while human
preference-based metrics focus on abstract quality assessments rather than
anatomical fidelity. To address this gap, we establish the first standards for
identifying human body distortions in AI-generated images and introduce
Distortion-5K, a comprehensive dataset comprising 4,700 annotated images of
normal and malformed human figures across diverse styles and distortion types.
Based on this dataset, we propose ViT-HD, a Vision Transformer-based model
tailored for detecting human body distortions in AI-generated images, which
outperforms state-of-the-art segmentation models and visual language models,
achieving an F1 score of 0.899 and IoU of 0.831 on distortion localization.
Additionally, we construct the Human Distortion Benchmark with 500
human-centric prompts to evaluate four popular T2I models using trained ViT-HD,
revealing that nearly 50\% of generated images contain distortions. This work
pioneers a systematic approach to evaluating anatomical accuracy in
AI-generated humans, offering tools to advance the fidelity of T2I models and
their real-world applicability. The Distortion-5K dataset, trained ViT-HD will
soon be released in our GitHub repository:
\href{https://github.com/TheRoadQaQ/Predicting-Distortion}{https://github.com/TheRoadQaQ/Predicting-Distortion}.
|
new_dataset
| 0.96378 |
2503.00814
|
Min Wang
|
Min Wang, Haisheng Li, Haoxuan Zhang, Xiaoqun Wu, Nan Li
|
PINN-MG: A physics-informed neural network for mesh generation
|
Accepted by Chinagraph2024 and recommended for publication in
Communications in Information and Systems
| null | null | null |
cs.CE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In numerical simulation, structured mesh generation often requires a lot of
time and manpower investment. The general scheme for structured quad mesh
generation is to find a mapping between the computational domain and the
physical domain. This mapping can be obtained by solving partial differential
equations. However, existing structured mesh generation methods are difficult
to ensure both efficiency and mesh quality. In this paper, we propose a
structured mesh generation method based on physics-informed neural network,
PINN-MG. It takes boundary curves as input and then utilizes an attention
network to capture the potential mapping between computational and physical
domains, generating structured meshes for the input physical domain. PINN-MG
introduces the Navier-Lam\'e equation in linear elastic as a partial
differential equation term in the loss function, ensuring that the neural
network conforms to the law of elastic body deformation when optimizing the
loss value. The training process of PINN-MG is completely unsupervised and does
not require any prior knowledge or datasets, which greatly reduces the previous
workload of producing structured mesh datasets. Experimental results show that
PINN-MG can generate higher quality structured quad meshes than other methods,
and has the advantages of traditional algebraic methods and differential
methods.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 09:43:42 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Wang",
"Min",
""
],
[
"Li",
"Haisheng",
""
],
[
"Zhang",
"Haoxuan",
""
],
[
"Wu",
"Xiaoqun",
""
],
[
"Li",
"Nan",
""
]
] |
TITLE: PINN-MG: A physics-informed neural network for mesh generation
ABSTRACT: In numerical simulation, structured mesh generation often requires a lot of
time and manpower investment. The general scheme for structured quad mesh
generation is to find a mapping between the computational domain and the
physical domain. This mapping can be obtained by solving partial differential
equations. However, existing structured mesh generation methods are difficult
to ensure both efficiency and mesh quality. In this paper, we propose a
structured mesh generation method based on physics-informed neural network,
PINN-MG. It takes boundary curves as input and then utilizes an attention
network to capture the potential mapping between computational and physical
domains, generating structured meshes for the input physical domain. PINN-MG
introduces the Navier-Lam\'e equation in linear elastic as a partial
differential equation term in the loss function, ensuring that the neural
network conforms to the law of elastic body deformation when optimizing the
loss value. The training process of PINN-MG is completely unsupervised and does
not require any prior knowledge or datasets, which greatly reduces the previous
workload of producing structured mesh datasets. Experimental results show that
PINN-MG can generate higher quality structured quad meshes than other methods,
and has the advantages of traditional algebraic methods and differential
methods.
|
no_new_dataset
| 0.949342 |
2503.00828
|
Yang He
|
Yalun Dai, Lingao Xiao, Ivor W. Tsang, Yang He
|
Training-Free Dataset Pruning for Instance Segmentation
|
Accepted by ICLR 2025
| null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Existing dataset pruning techniques primarily focus on classification tasks,
limiting their applicability to more complex and practical tasks like instance
segmentation. Instance segmentation presents three key challenges: pixel-level
annotations, instance area variations, and class imbalances, which
significantly complicate dataset pruning efforts. Directly adapting existing
classification-based pruning methods proves ineffective due to their reliance
on time-consuming model training process. To address this, we propose a novel
Training-Free Dataset Pruning (TFDP) method for instance segmentation.
Specifically, we leverage shape and class information from image annotations to
design a Shape Complexity Score (SCS), refining it into a Scale-Invariant
(SI-SCS) and Class-Balanced (CB-SCS) versions to address instance area
variations and class imbalances, all without requiring model training. We
achieve state-of-the-art results on VOC 2012, Cityscapes, and COCO datasets,
generalizing well across CNN and Transformer architectures. Remarkably, our
approach accelerates the pruning process by an average of 1349$\times$ on COCO
compared to the adapted baselines. Source code is available at:
https://github.com/he-y/dataset-pruning-for-instance-segmentation
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 10:05:59 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Dai",
"Yalun",
""
],
[
"Xiao",
"Lingao",
""
],
[
"Tsang",
"Ivor W.",
""
],
[
"He",
"Yang",
""
]
] |
TITLE: Training-Free Dataset Pruning for Instance Segmentation
ABSTRACT: Existing dataset pruning techniques primarily focus on classification tasks,
limiting their applicability to more complex and practical tasks like instance
segmentation. Instance segmentation presents three key challenges: pixel-level
annotations, instance area variations, and class imbalances, which
significantly complicate dataset pruning efforts. Directly adapting existing
classification-based pruning methods proves ineffective due to their reliance
on time-consuming model training process. To address this, we propose a novel
Training-Free Dataset Pruning (TFDP) method for instance segmentation.
Specifically, we leverage shape and class information from image annotations to
design a Shape Complexity Score (SCS), refining it into a Scale-Invariant
(SI-SCS) and Class-Balanced (CB-SCS) versions to address instance area
variations and class imbalances, all without requiring model training. We
achieve state-of-the-art results on VOC 2012, Cityscapes, and COCO datasets,
generalizing well across CNN and Transformer architectures. Remarkably, our
approach accelerates the pruning process by an average of 1349$\times$ on COCO
compared to the adapted baselines. Source code is available at:
https://github.com/he-y/dataset-pruning-for-instance-segmentation
|
no_new_dataset
| 0.949856 |
2503.00841
|
Jiaxin Shen
|
Jiaxin Shen, Jinan Xu, Huiqi Hu, Luyi Lin, Fei Zheng, Guoyang Ma,
Fandong Meng, Jie Zhou, Wenjuan Han
|
A Law Reasoning Benchmark for LLM with Tree-Organized Structures
including Factum Probandum, Evidence and Experiences
|
20 pages, 13 figures
| null | null | null |
cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
While progress has been made in legal applications, law reasoning, crucial
for fair adjudication, remains unexplored. We propose a transparent law
reasoning schema enriched with hierarchical factum probandum, evidence, and
implicit experience, enabling public scrutiny and preventing bias. Inspired by
this schema, we introduce the challenging task, which takes a textual case
description and outputs a hierarchical structure justifying the final decision.
We also create the first crowd-sourced dataset for this task, enabling
comprehensive evaluation. Simultaneously, we propose an agent framework that
employs a comprehensive suite of legal analysis tools to address the challenge
task. This benchmark paves the way for transparent and accountable AI-assisted
law reasoning in the ``Intelligent Court''.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 10:26:54 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Shen",
"Jiaxin",
""
],
[
"Xu",
"Jinan",
""
],
[
"Hu",
"Huiqi",
""
],
[
"Lin",
"Luyi",
""
],
[
"Zheng",
"Fei",
""
],
[
"Ma",
"Guoyang",
""
],
[
"Meng",
"Fandong",
""
],
[
"Zhou",
"Jie",
""
],
[
"Han",
"Wenjuan",
""
]
] |
TITLE: A Law Reasoning Benchmark for LLM with Tree-Organized Structures
including Factum Probandum, Evidence and Experiences
ABSTRACT: While progress has been made in legal applications, law reasoning, crucial
for fair adjudication, remains unexplored. We propose a transparent law
reasoning schema enriched with hierarchical factum probandum, evidence, and
implicit experience, enabling public scrutiny and preventing bias. Inspired by
this schema, we introduce the challenging task, which takes a textual case
description and outputs a hierarchical structure justifying the final decision.
We also create the first crowd-sourced dataset for this task, enabling
comprehensive evaluation. Simultaneously, we propose an agent framework that
employs a comprehensive suite of legal analysis tools to address the challenge
task. This benchmark paves the way for transparent and accountable AI-assisted
law reasoning in the ``Intelligent Court''.
|
new_dataset
| 0.954052 |
2503.00845
|
Miao Peng
|
Miao Peng, Nuo Chen, Zongrui Suo, Jia Li
|
Rewarding Graph Reasoning Process makes LLMs more Generalized Reasoners
| null | null | null | null |
cs.CL cs.AI cs.LG
|
http://creativecommons.org/publicdomain/zero/1.0/
|
Despite significant advancements in Large Language Models (LLMs), developing
advanced reasoning capabilities in LLMs remains a key challenge. Process Reward
Models (PRMs) have demonstrated exceptional promise in enhancing reasoning by
providing step-wise feedback, particularly in the context of mathematical
reasoning. However, their application to broader reasoning domains remains
understudied, largely due to the high costs associated with manually creating
step-level supervision. In this work, we explore the potential of PRMs in graph
reasoning problems - a domain that demands sophisticated multi-step reasoning
and offers opportunities for automated step-level data generation using
established graph algorithms. We introduce GraphSILO, the largest dataset for
graph reasoning problems with fine-grained step-wise labels, built using
automated Task-oriented Trajectories and Monte Carlo Tree Search (MCTS) to
generate detailed reasoning steps with step-wise labels. Building upon this
dataset, we train GraphPRM, the first PRM designed for graph reasoning
problems, and evaluate its effectiveness in two key settings: inference-time
scaling and reinforcement learning via Direct Preference Optimization (DPO).
Experimental results show that GraphPRM significantly improves LLM performance
across 13 graph reasoning tasks, delivering a 9% gain for Qwen2.5-7B and
demonstrating transferability to new graph reasoning datasets and new reasoning
domains like mathematical problem-solving. Notably, GraphPRM enhances LLM
performance on GSM8K and Math500, underscoring the cross-domain applicability
of graph-based reasoning rewards. Our findings highlight the potential of PRMs
in advancing reasoning across diverse domains, paving the way for more
versatile and effective LLMs.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 10:39:40 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Peng",
"Miao",
""
],
[
"Chen",
"Nuo",
""
],
[
"Suo",
"Zongrui",
""
],
[
"Li",
"Jia",
""
]
] |
TITLE: Rewarding Graph Reasoning Process makes LLMs more Generalized Reasoners
ABSTRACT: Despite significant advancements in Large Language Models (LLMs), developing
advanced reasoning capabilities in LLMs remains a key challenge. Process Reward
Models (PRMs) have demonstrated exceptional promise in enhancing reasoning by
providing step-wise feedback, particularly in the context of mathematical
reasoning. However, their application to broader reasoning domains remains
understudied, largely due to the high costs associated with manually creating
step-level supervision. In this work, we explore the potential of PRMs in graph
reasoning problems - a domain that demands sophisticated multi-step reasoning
and offers opportunities for automated step-level data generation using
established graph algorithms. We introduce GraphSILO, the largest dataset for
graph reasoning problems with fine-grained step-wise labels, built using
automated Task-oriented Trajectories and Monte Carlo Tree Search (MCTS) to
generate detailed reasoning steps with step-wise labels. Building upon this
dataset, we train GraphPRM, the first PRM designed for graph reasoning
problems, and evaluate its effectiveness in two key settings: inference-time
scaling and reinforcement learning via Direct Preference Optimization (DPO).
Experimental results show that GraphPRM significantly improves LLM performance
across 13 graph reasoning tasks, delivering a 9% gain for Qwen2.5-7B and
demonstrating transferability to new graph reasoning datasets and new reasoning
domains like mathematical problem-solving. Notably, GraphPRM enhances LLM
performance on GSM8K and Math500, underscoring the cross-domain applicability
of graph-based reasoning rewards. Our findings highlight the potential of PRMs
in advancing reasoning across diverse domains, paving the way for more
versatile and effective LLMs.
|
new_dataset
| 0.956997 |
2503.00848
|
Bocheng Li
|
BoCheng Li, WenJuan Zhang, Bing Zhang, YiLing Yao, YaNing Wang
|
PSRGS:Progressive Spectral Residual of 3D Gaussian for High-Frequency
Recovery
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
3D Gaussian Splatting (3D GS) achieves impressive results in novel view
synthesis for small, single-object scenes through Gaussian ellipsoid
initialization and adaptive density control. However, when applied to
large-scale remote sensing scenes, 3D GS faces challenges: the point clouds
generated by Structure-from-Motion (SfM) are often sparse, and the inherent
smoothing behavior of 3D GS leads to over-reconstruction in high-frequency
regions, where have detailed textures and color variations. This results in the
generation of large, opaque Gaussian ellipsoids that cause gradient artifacts.
Moreover, the simultaneous optimization of both geometry and texture may lead
to densification of Gaussian ellipsoids at incorrect geometric locations,
resulting in artifacts in other views. To address these issues, we propose
PSRGS, a progressive optimization scheme based on spectral residual maps.
Specifically, we create a spectral residual significance map to separate
low-frequency and high-frequency regions. In the low-frequency region, we apply
depth-aware and depth-smooth losses to initialize the scene geometry with low
threshold. For the high-frequency region, we use gradient features with higher
threshold to split and clone ellipsoids, refining the scene. The sampling rate
is determined by feature responses and gradient loss. Finally, we introduce a
pre-trained network that jointly computes perceptual loss from multiple views,
ensuring accurate restoration of high-frequency details in both Gaussian
ellipsoids geometry and color. We conduct experiments on multiple datasets to
assess the effectiveness of our method, which demonstrates competitive
rendering quality, especially in recovering texture details in high-frequency
regions.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 10:52:46 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Li",
"BoCheng",
""
],
[
"Zhang",
"WenJuan",
""
],
[
"Zhang",
"Bing",
""
],
[
"Yao",
"YiLing",
""
],
[
"Wang",
"YaNing",
""
]
] |
TITLE: PSRGS:Progressive Spectral Residual of 3D Gaussian for High-Frequency
Recovery
ABSTRACT: 3D Gaussian Splatting (3D GS) achieves impressive results in novel view
synthesis for small, single-object scenes through Gaussian ellipsoid
initialization and adaptive density control. However, when applied to
large-scale remote sensing scenes, 3D GS faces challenges: the point clouds
generated by Structure-from-Motion (SfM) are often sparse, and the inherent
smoothing behavior of 3D GS leads to over-reconstruction in high-frequency
regions, where have detailed textures and color variations. This results in the
generation of large, opaque Gaussian ellipsoids that cause gradient artifacts.
Moreover, the simultaneous optimization of both geometry and texture may lead
to densification of Gaussian ellipsoids at incorrect geometric locations,
resulting in artifacts in other views. To address these issues, we propose
PSRGS, a progressive optimization scheme based on spectral residual maps.
Specifically, we create a spectral residual significance map to separate
low-frequency and high-frequency regions. In the low-frequency region, we apply
depth-aware and depth-smooth losses to initialize the scene geometry with low
threshold. For the high-frequency region, we use gradient features with higher
threshold to split and clone ellipsoids, refining the scene. The sampling rate
is determined by feature responses and gradient loss. Finally, we introduce a
pre-trained network that jointly computes perceptual loss from multiple views,
ensuring accurate restoration of high-frequency details in both Gaussian
ellipsoids geometry and color. We conduct experiments on multiple datasets to
assess the effectiveness of our method, which demonstrates competitive
rendering quality, especially in recovering texture details in high-frequency
regions.
|
no_new_dataset
| 0.950824 |
2503.00853
|
Arnold Wiliem
|
Rui Yi Yong and Samuel Picosson and Arnold Wiliem
|
MTReD: 3D Reconstruction Dataset for Fly-over Videos of Maritime Domain
|
WACV Workshop 2025 - 3rd Workshop on Maritime Computer Vision
(MaCVI2025)
|
3rd Workshop on Maritime Computer Vision, WACV 2025 Workshop
| null | null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This work tackles 3D scene reconstruction for a video fly-over perspective
problem in the maritime domain, with a specific emphasis on geometrically and
visually sound reconstructions. This will allow for downstream tasks such as
segmentation, navigation, and localization. To our knowledge, there is no
dataset available in this domain. As such, we propose a novel maritime 3D scene
reconstruction benchmarking dataset, named as MTReD (Maritime Three-Dimensional
Reconstruction Dataset). The MTReD comprises 19 fly-over videos curated from
the Internet containing ships, islands, and coastlines. As the task is aimed
towards geometrical consistency and visual completeness, the dataset uses two
metrics: (1) Reprojection error; and (2) Perception based metrics. We find that
existing perception-based metrics, such as Learned Perceptual Image Patch
Similarity (LPIPS), do not appropriately measure the completeness of a
reconstructed image. Thus, we propose a novel semantic similarity metric
utilizing DINOv2 features coined DiFPS (DinoV2 Features Perception Similarity).
We perform initial evaluation on two baselines: (1) Structured from Motion
(SfM) through Colmap; and (2) the recent state-of-the-art MASt3R model. We find
that the reconstructed scenes by MASt3R have higher reprojection errors, but
superior perception based metric scores. To this end, some pre-processing
methods are explored, and we find a pre-processing method which improves both
the reprojection error and perception-based score. We envisage our proposed
MTReD to stimulate further research in these directions. The dataset and all
the code will be made available in https://github.com/RuiYiYong/MTReD.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 11:10:34 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Yong",
"Rui Yi",
""
],
[
"Picosson",
"Samuel",
""
],
[
"Wiliem",
"Arnold",
""
]
] |
TITLE: MTReD: 3D Reconstruction Dataset for Fly-over Videos of Maritime Domain
ABSTRACT: This work tackles 3D scene reconstruction for a video fly-over perspective
problem in the maritime domain, with a specific emphasis on geometrically and
visually sound reconstructions. This will allow for downstream tasks such as
segmentation, navigation, and localization. To our knowledge, there is no
dataset available in this domain. As such, we propose a novel maritime 3D scene
reconstruction benchmarking dataset, named as MTReD (Maritime Three-Dimensional
Reconstruction Dataset). The MTReD comprises 19 fly-over videos curated from
the Internet containing ships, islands, and coastlines. As the task is aimed
towards geometrical consistency and visual completeness, the dataset uses two
metrics: (1) Reprojection error; and (2) Perception based metrics. We find that
existing perception-based metrics, such as Learned Perceptual Image Patch
Similarity (LPIPS), do not appropriately measure the completeness of a
reconstructed image. Thus, we propose a novel semantic similarity metric
utilizing DINOv2 features coined DiFPS (DinoV2 Features Perception Similarity).
We perform initial evaluation on two baselines: (1) Structured from Motion
(SfM) through Colmap; and (2) the recent state-of-the-art MASt3R model. We find
that the reconstructed scenes by MASt3R have higher reprojection errors, but
superior perception based metric scores. To this end, some pre-processing
methods are explored, and we find a pre-processing method which improves both
the reprojection error and perception-based score. We envisage our proposed
MTReD to stimulate further research in these directions. The dataset and all
the code will be made available in https://github.com/RuiYiYong/MTReD.
|
new_dataset
| 0.973368 |
2503.00854
|
Tai Le Quy
|
Tai Le Quy, Long Le Thanh, Lan Luong Thi Hong, Frank Hopfgartner
|
FACROC: a fairness measure for FAir Clustering through ROC curves
|
Accepted to Special Session: Data Science: Foundations and
Applications (DSFA), PAKDD 2025
| null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Fair clustering has attracted remarkable attention from the research
community. Many fairness measures for clustering have been proposed; however,
they do not take into account the clustering quality w.r.t. the values of the
protected attribute. In this paper, we introduce a new visual-based fairness
measure for fair clustering through ROC curves, namely FACROC. This fairness
measure employs AUCC as a measure of clustering quality and then computes the
difference in the corresponding ROC curves for each value of the protected
attribute. Experimental results on several popular datasets for fairness-aware
machine learning and well-known (fair) clustering models show that FACROC is a
beneficial method for visually evaluating the fairness of clustering models.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 11:11:34 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Quy",
"Tai Le",
""
],
[
"Thanh",
"Long Le",
""
],
[
"Hong",
"Lan Luong Thi",
""
],
[
"Hopfgartner",
"Frank",
""
]
] |
TITLE: FACROC: a fairness measure for FAir Clustering through ROC curves
ABSTRACT: Fair clustering has attracted remarkable attention from the research
community. Many fairness measures for clustering have been proposed; however,
they do not take into account the clustering quality w.r.t. the values of the
protected attribute. In this paper, we introduce a new visual-based fairness
measure for fair clustering through ROC curves, namely FACROC. This fairness
measure employs AUCC as a measure of clustering quality and then computes the
difference in the corresponding ROC curves for each value of the protected
attribute. Experimental results on several popular datasets for fairness-aware
machine learning and well-known (fair) clustering models show that FACROC is a
beneficial method for visually evaluating the fairness of clustering models.
|
no_new_dataset
| 0.953405 |
2503.00865
|
Yiran Zhao
|
Yiran Zhao, Chaoqun Liu, Yue Deng, Jiahao Ying, Mahani Aljunied,
Zhaodonghui Li, Lidong Bing, Hou Pong Chan, Yu Rong, Deli Zhao, Wenxuan Zhang
|
Babel: Open Multilingual Large Language Models Serving Over 90% of
Global Speakers
| null | null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Large language models (LLMs) have revolutionized natural language processing
(NLP), yet open-source multilingual LLMs remain scarce, with existing models
often limited in language coverage. Such models typically prioritize
well-resourced languages, while widely spoken but under-resourced languages are
often overlooked. To address this disparity, we introduce $\texttt{Babel}$, an
open multilingual LLM that covers the top 25 languages by number of speakers,
supports over 90% of the global population, and includes many languages
neglected by other open multilingual LLMs. Unlike traditional continue
pretraining approaches, Babel expands its parameter count through a layer
extension technique that elevates Babel's performance ceiling. We introduce two
variants: $\texttt{Babel-9B}$, designed for efficient inference and
fine-tuning, and $\texttt{Babel-83B}$, which sets a new standard for open
multilingual LLMs. Extensive evaluations on multilingual tasks demonstrate its
superior performance compared to open LLMs of comparable size. In addition,
using open-source supervised fine-tuning datasets, Babel achieves remarkable
performance, with Babel-9B-Chat leading among 10B-sized LLMs and Babel-83B-Chat
setting a new standard for multilingual tasks, reaching the same level of
commercial models.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 11:53:55 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhao",
"Yiran",
""
],
[
"Liu",
"Chaoqun",
""
],
[
"Deng",
"Yue",
""
],
[
"Ying",
"Jiahao",
""
],
[
"Aljunied",
"Mahani",
""
],
[
"Li",
"Zhaodonghui",
""
],
[
"Bing",
"Lidong",
""
],
[
"Chan",
"Hou Pong",
""
],
[
"Rong",
"Yu",
""
],
[
"Zhao",
"Deli",
""
],
[
"Zhang",
"Wenxuan",
""
]
] |
TITLE: Babel: Open Multilingual Large Language Models Serving Over 90% of
Global Speakers
ABSTRACT: Large language models (LLMs) have revolutionized natural language processing
(NLP), yet open-source multilingual LLMs remain scarce, with existing models
often limited in language coverage. Such models typically prioritize
well-resourced languages, while widely spoken but under-resourced languages are
often overlooked. To address this disparity, we introduce $\texttt{Babel}$, an
open multilingual LLM that covers the top 25 languages by number of speakers,
supports over 90% of the global population, and includes many languages
neglected by other open multilingual LLMs. Unlike traditional continue
pretraining approaches, Babel expands its parameter count through a layer
extension technique that elevates Babel's performance ceiling. We introduce two
variants: $\texttt{Babel-9B}$, designed for efficient inference and
fine-tuning, and $\texttt{Babel-83B}$, which sets a new standard for open
multilingual LLMs. Extensive evaluations on multilingual tasks demonstrate its
superior performance compared to open LLMs of comparable size. In addition,
using open-source supervised fine-tuning datasets, Babel achieves remarkable
performance, with Babel-9B-Chat leading among 10B-sized LLMs and Babel-83B-Chat
setting a new standard for multilingual tasks, reaching the same level of
commercial models.
|
no_new_dataset
| 0.947284 |
2503.00867
|
Alexios Gidiotis
|
Petros Stylianos Giouroukis, Alexios Gidiotis, Grigorios Tsoumakas
|
DUAL: Diversity and Uncertainty Active Learning for Text Summarization
| null | null | null | null |
cs.CL
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
With the rise of large language models, neural text summarization has
advanced significantly in recent years. However, even state-of-the-art models
continue to rely heavily on high-quality human-annotated data for training and
evaluation. Active learning is frequently used as an effective way to collect
such datasets, especially when annotation resources are scarce. Active learning
methods typically prioritize either uncertainty or diversity but have shown
limited effectiveness in summarization, often being outperformed by random
sampling. We present Diversity and Uncertainty Active Learning (DUAL), a novel
algorithm that combines uncertainty and diversity to iteratively select and
annotate samples that are both representative of the data distribution and
challenging for the current model. DUAL addresses the selection of noisy
samples in uncertainty-based methods and the limited exploration scope of
diversity-based methods. Through extensive experiments with different
summarization models and benchmark datasets, we demonstrate that DUAL
consistently matches or outperforms the best performing strategies. Using
visualizations and quantitative metrics, we provide valuable insights into the
effectiveness and robustness of different active learning strategies, in an
attempt to understand why these strategies haven't performed consistently in
text summarization. Finally, we show that DUAL strikes a good balance between
diversity and robustness.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 12:06:16 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Giouroukis",
"Petros Stylianos",
""
],
[
"Gidiotis",
"Alexios",
""
],
[
"Tsoumakas",
"Grigorios",
""
]
] |
TITLE: DUAL: Diversity and Uncertainty Active Learning for Text Summarization
ABSTRACT: With the rise of large language models, neural text summarization has
advanced significantly in recent years. However, even state-of-the-art models
continue to rely heavily on high-quality human-annotated data for training and
evaluation. Active learning is frequently used as an effective way to collect
such datasets, especially when annotation resources are scarce. Active learning
methods typically prioritize either uncertainty or diversity but have shown
limited effectiveness in summarization, often being outperformed by random
sampling. We present Diversity and Uncertainty Active Learning (DUAL), a novel
algorithm that combines uncertainty and diversity to iteratively select and
annotate samples that are both representative of the data distribution and
challenging for the current model. DUAL addresses the selection of noisy
samples in uncertainty-based methods and the limited exploration scope of
diversity-based methods. Through extensive experiments with different
summarization models and benchmark datasets, we demonstrate that DUAL
consistently matches or outperforms the best performing strategies. Using
visualizations and quantitative metrics, we provide valuable insights into the
effectiveness and robustness of different active learning strategies, in an
attempt to understand why these strategies haven't performed consistently in
text summarization. Finally, we show that DUAL strikes a good balance between
diversity and robustness.
|
no_new_dataset
| 0.94743 |
2503.00871
|
Kota Nakamura
|
Kota Nakamura, Koki Kawabata, Shungo Tanaka, Yasuko Matsubara, Yasushi
Sakurai
|
CyberCScope: Mining Skewed Tensor Streams and Online Anomaly Detection
in Cybersecurity Systems
|
Accepted by WWW 2025 short research paper
| null |
10.1145/3701716.3715476
| null |
cs.LG cs.AI cs.CR
|
http://creativecommons.org/licenses/by/4.0/
|
Cybersecurity systems are continuously producing a huge number of
time-stamped events in the form of high-order tensors, such as {count; time,
port, flow duration, packet size, . . . }, and so how can we detect
anomalies/intrusions in real time? How can we identify multiple types of
intrusions and capture their characteristic behaviors? The tensor data consists
of categorical and continuous attributes and the data distributions of
continuous attributes typically exhibit skew. These data properties require
handling skewed infinite and finite dimensional spaces simultaneously. In this
paper, we propose a novel streaming method, namely CyberCScope. The method
effectively decomposes incoming tensors into major trends while explicitly
distinguishing between categorical and skewed continuous attributes. To our
knowledge, it is the first to compute hybrid skewed infinite and finite
dimensional decomposition. Based on this decomposition, it streamingly finds
distinct time-evolving patterns, enabling the detection of multiple types of
anomalies. Extensive experiments on large-scale real datasets demonstrate that
CyberCScope detects various intrusions with higher accuracy than
state-of-the-art baselines while providing meaningful summaries for the
intrusions that occur in practice.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 12:17:24 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Nakamura",
"Kota",
""
],
[
"Kawabata",
"Koki",
""
],
[
"Tanaka",
"Shungo",
""
],
[
"Matsubara",
"Yasuko",
""
],
[
"Sakurai",
"Yasushi",
""
]
] |
TITLE: CyberCScope: Mining Skewed Tensor Streams and Online Anomaly Detection
in Cybersecurity Systems
ABSTRACT: Cybersecurity systems are continuously producing a huge number of
time-stamped events in the form of high-order tensors, such as {count; time,
port, flow duration, packet size, . . . }, and so how can we detect
anomalies/intrusions in real time? How can we identify multiple types of
intrusions and capture their characteristic behaviors? The tensor data consists
of categorical and continuous attributes and the data distributions of
continuous attributes typically exhibit skew. These data properties require
handling skewed infinite and finite dimensional spaces simultaneously. In this
paper, we propose a novel streaming method, namely CyberCScope. The method
effectively decomposes incoming tensors into major trends while explicitly
distinguishing between categorical and skewed continuous attributes. To our
knowledge, it is the first to compute hybrid skewed infinite and finite
dimensional decomposition. Based on this decomposition, it streamingly finds
distinct time-evolving patterns, enabling the detection of multiple types of
anomalies. Extensive experiments on large-scale real datasets demonstrate that
CyberCScope detects various intrusions with higher accuracy than
state-of-the-art baselines while providing meaningful summaries for the
intrusions that occur in practice.
|
no_new_dataset
| 0.949201 |
2503.00877
|
Dilfira Kudrat
|
Dilfira Kudrat, Zongxia Xie, Yanru Sun, Tianyu Jia, Qinghua Hu
|
Patch-wise Structural Loss for Time Series Forecasting
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Time-series forecasting has gained significant attention in machine learning
due to its crucial role in various domains. However, most existing forecasting
models rely heavily on point-wise loss functions like Mean Square Error, which
treat each time step independently and neglect the structural dependencies
inherent in time series data, making it challenging to capture complex temporal
patterns accurately. To address these challenges, we propose a novel Patch-wise
Structural (PS) loss, designed to enhance structural alignment by comparing
time series at the patch level. Through leveraging local statistical
properties, such as correlation, variance, and mean, PS loss captures nuanced
structural discrepancies overlooked by traditional point-wise losses.
Furthermore, it integrates seamlessly with point-wise loss, simultaneously
addressing local structural inconsistencies and individual time-step errors. PS
loss establishes a novel benchmark for accurately modeling complex time series
data and provides a new perspective on time series loss function design.
Extensive experiments demonstrate that PS loss significantly improves the
performance of state-of-the-art models across diverse real-world datasets.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 12:36:15 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Kudrat",
"Dilfira",
""
],
[
"Xie",
"Zongxia",
""
],
[
"Sun",
"Yanru",
""
],
[
"Jia",
"Tianyu",
""
],
[
"Hu",
"Qinghua",
""
]
] |
TITLE: Patch-wise Structural Loss for Time Series Forecasting
ABSTRACT: Time-series forecasting has gained significant attention in machine learning
due to its crucial role in various domains. However, most existing forecasting
models rely heavily on point-wise loss functions like Mean Square Error, which
treat each time step independently and neglect the structural dependencies
inherent in time series data, making it challenging to capture complex temporal
patterns accurately. To address these challenges, we propose a novel Patch-wise
Structural (PS) loss, designed to enhance structural alignment by comparing
time series at the patch level. Through leveraging local statistical
properties, such as correlation, variance, and mean, PS loss captures nuanced
structural discrepancies overlooked by traditional point-wise losses.
Furthermore, it integrates seamlessly with point-wise loss, simultaneously
addressing local structural inconsistencies and individual time-step errors. PS
loss establishes a novel benchmark for accurately modeling complex time series
data and provides a new perspective on time series loss function design.
Extensive experiments demonstrate that PS loss significantly improves the
performance of state-of-the-art models across diverse real-world datasets.
|
no_new_dataset
| 0.953013 |
2503.00884
|
Rundong He
|
Rundong He, Yicong Dong, Lanzhe Guo, Yilong Yin, Tailin Wu
|
Re-Evaluating the Impact of Unseen-Class Unlabeled Data on
Semi-Supervised Learning Model
|
Published as a conference paper at ICLR 2025
| null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Semi-supervised learning (SSL) effectively leverages unlabeled data and has
been proven successful across various fields. Current safe SSL methods believe
that unseen classes in unlabeled data harm the performance of SSL models.
However, previous methods for assessing the impact of unseen classes on SSL
model performance are flawed. They fix the size of the unlabeled dataset and
adjust the proportion of unseen classes within the unlabeled data to assess the
impact. This process contravenes the principle of controlling variables.
Adjusting the proportion of unseen classes in unlabeled data alters the
proportion of seen classes, meaning the decreased classification performance of
seen classes may not be due to an increase in unseen class samples in the
unlabeled data, but rather a decrease in seen class samples. Thus, the prior
flawed assessment standard that ``unseen classes in unlabeled data can damage
SSL model performance" may not always hold true. This paper strictly adheres to
the principle of controlling variables, maintaining the proportion of seen
classes in unlabeled data while only changing the unseen classes across five
critical dimensions, to investigate their impact on SSL models from global
robustness and local robustness. Experiments demonstrate that unseen classes in
unlabeled data do not necessarily impair the performance of SSL models; in
fact, under certain conditions, unseen classes may even enhance them.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 13:06:00 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"He",
"Rundong",
""
],
[
"Dong",
"Yicong",
""
],
[
"Guo",
"Lanzhe",
""
],
[
"Yin",
"Yilong",
""
],
[
"Wu",
"Tailin",
""
]
] |
TITLE: Re-Evaluating the Impact of Unseen-Class Unlabeled Data on
Semi-Supervised Learning Model
ABSTRACT: Semi-supervised learning (SSL) effectively leverages unlabeled data and has
been proven successful across various fields. Current safe SSL methods believe
that unseen classes in unlabeled data harm the performance of SSL models.
However, previous methods for assessing the impact of unseen classes on SSL
model performance are flawed. They fix the size of the unlabeled dataset and
adjust the proportion of unseen classes within the unlabeled data to assess the
impact. This process contravenes the principle of controlling variables.
Adjusting the proportion of unseen classes in unlabeled data alters the
proportion of seen classes, meaning the decreased classification performance of
seen classes may not be due to an increase in unseen class samples in the
unlabeled data, but rather a decrease in seen class samples. Thus, the prior
flawed assessment standard that ``unseen classes in unlabeled data can damage
SSL model performance" may not always hold true. This paper strictly adheres to
the principle of controlling variables, maintaining the proportion of seen
classes in unlabeled data while only changing the unseen classes across five
critical dimensions, to investigate their impact on SSL models from global
robustness and local robustness. Experiments demonstrate that unseen classes in
unlabeled data do not necessarily impair the performance of SSL models; in
fact, under certain conditions, unseen classes may even enhance them.
|
no_new_dataset
| 0.946349 |
2503.00898
|
Nico Reeb
|
Nico Reeb, Javier Lopez-Randulfe, Robin Dietrich and Alois C. Knoll
|
Range and Angle Estimation with Spiking Neural Resonators for FMCW Radar
| null | null | null | null |
cs.NE
|
http://creativecommons.org/licenses/by/4.0/
|
Automotive radar systems face the challenge of managing high sampling rates
and large data bandwidth while complying with stringent real-time and energy
efficiency requirements. The growing complexity of autonomous vehicles further
intensifies these requirements. Neuromorphic computing offers promising
solutions because of its inherent energy efficiency and parallel processing
capacity. This research presents a novel spiking neuron model for signal
processing of frequency-modulated continuous wave (FMCW) radars that
outperforms the state-of-the-art spectrum analysis algorithms in latency and
data bandwidth. These spiking neural resonators are based on the
resonate-and-fire neuron model and optimized to dynamically process raw radar
data while simultaneously emitting an output in the form of spikes. We designed
the first neuromorphic neural network consisting of these spiking neural
resonators that estimates range and angle from FMCW radar data. We evaluated
the range-angle maps on simulated datasets covering multiple scenarios and
compared the results with a state-of-the-art pipeline for radar processing. The
proposed neuron model significantly reduces the processing latency compared to
traditional frequency analysis algorithms, such as the Fourier transformation
(FT), which needs to sample and store entire data frames before processing. The
evaluations demonstrate that these spiking neural resonators achieve
state-of-the-art detection accuracy while emitting spikes simultaneously to
processing and transmitting only 0.02 % of the data compared to a float-32 FT.
The results showcase the potential for neuromorphic signal processing for FMCW
radar systems and pave the way for designing neuromorphic radar sensors.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 13:51:03 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Reeb",
"Nico",
""
],
[
"Lopez-Randulfe",
"Javier",
""
],
[
"Dietrich",
"Robin",
""
],
[
"Knoll",
"Alois C.",
""
]
] |
TITLE: Range and Angle Estimation with Spiking Neural Resonators for FMCW Radar
ABSTRACT: Automotive radar systems face the challenge of managing high sampling rates
and large data bandwidth while complying with stringent real-time and energy
efficiency requirements. The growing complexity of autonomous vehicles further
intensifies these requirements. Neuromorphic computing offers promising
solutions because of its inherent energy efficiency and parallel processing
capacity. This research presents a novel spiking neuron model for signal
processing of frequency-modulated continuous wave (FMCW) radars that
outperforms the state-of-the-art spectrum analysis algorithms in latency and
data bandwidth. These spiking neural resonators are based on the
resonate-and-fire neuron model and optimized to dynamically process raw radar
data while simultaneously emitting an output in the form of spikes. We designed
the first neuromorphic neural network consisting of these spiking neural
resonators that estimates range and angle from FMCW radar data. We evaluated
the range-angle maps on simulated datasets covering multiple scenarios and
compared the results with a state-of-the-art pipeline for radar processing. The
proposed neuron model significantly reduces the processing latency compared to
traditional frequency analysis algorithms, such as the Fourier transformation
(FT), which needs to sample and store entire data frames before processing. The
evaluations demonstrate that these spiking neural resonators achieve
state-of-the-art detection accuracy while emitting spikes simultaneously to
processing and transmitting only 0.02 % of the data compared to a float-32 FT.
The results showcase the potential for neuromorphic signal processing for FMCW
radar systems and pave the way for designing neuromorphic radar sensors.
|
no_new_dataset
| 0.956391 |
2503.00900
|
Qiong Zhang
|
Jing Peng and Meiqi Yang and Qiong Zhang and Xiaoxiao Li
|
S4M: S4 for multivariate time series forecasting with Missing values
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Multivariate time series data play a pivotal role in a wide range of
real-world applications. However, the presence of block missing data introduces
significant challenges, often compromising the performance of predictive
models. Traditional two-step approaches, which first impute missing values and
then perform forecasting, are prone to error accumulation, particularly in
complex multivariate settings characterized by high missing ratios and
intricate dependency structures. In this work, we introduce S4M, an end-to-end
time series forecasting framework that seamlessly integrates missing data
handling into the Structured State Space Sequence (S4) model architecture.
Unlike conventional methods that treat imputation as a separate preprocessing
step, S4M leverages the latent space of S4 models to directly recognize and
represent missing data patterns, thereby more effectively capturing the
underlying temporal and multivariate dependencies. Our framework comprises two
key components: the Adaptive Temporal Prototype Mapper (ATPM) and the
Missing-Aware Dual Stream S4 (MDS-S4). The ATPM employs a prototype bank to
derive robust and informative representations from historical data patterns,
while the MDS-S4 processes these representations alongside missingness masks as
dual input streams to enable accurate forecasting. Through extensive empirical
evaluations on diverse real-world datasets, we demonstrate that S4M
consistently achieves state-of-the-art performance. These results underscore
the efficacy of our integrated approach in handling missing data, showcasing
its robustness and superiority over traditional imputation-based methods. Our
findings highlight the potential of S4M to advance reliable time series
forecasting in practical applications, offering a promising direction for
future research and deployment. Code is available at
https://github.com/WINTERWEEL/S4M.git.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 13:59:59 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Peng",
"Jing",
""
],
[
"Yang",
"Meiqi",
""
],
[
"Zhang",
"Qiong",
""
],
[
"Li",
"Xiaoxiao",
""
]
] |
TITLE: S4M: S4 for multivariate time series forecasting with Missing values
ABSTRACT: Multivariate time series data play a pivotal role in a wide range of
real-world applications. However, the presence of block missing data introduces
significant challenges, often compromising the performance of predictive
models. Traditional two-step approaches, which first impute missing values and
then perform forecasting, are prone to error accumulation, particularly in
complex multivariate settings characterized by high missing ratios and
intricate dependency structures. In this work, we introduce S4M, an end-to-end
time series forecasting framework that seamlessly integrates missing data
handling into the Structured State Space Sequence (S4) model architecture.
Unlike conventional methods that treat imputation as a separate preprocessing
step, S4M leverages the latent space of S4 models to directly recognize and
represent missing data patterns, thereby more effectively capturing the
underlying temporal and multivariate dependencies. Our framework comprises two
key components: the Adaptive Temporal Prototype Mapper (ATPM) and the
Missing-Aware Dual Stream S4 (MDS-S4). The ATPM employs a prototype bank to
derive robust and informative representations from historical data patterns,
while the MDS-S4 processes these representations alongside missingness masks as
dual input streams to enable accurate forecasting. Through extensive empirical
evaluations on diverse real-world datasets, we demonstrate that S4M
consistently achieves state-of-the-art performance. These results underscore
the efficacy of our integrated approach in handling missing data, showcasing
its robustness and superiority over traditional imputation-based methods. Our
findings highlight the potential of S4M to advance reliable time series
forecasting in practical applications, offering a promising direction for
future research and deployment. Code is available at
https://github.com/WINTERWEEL/S4M.git.
|
no_new_dataset
| 0.947478 |
2503.00908
|
Ziyuan Yang
|
Ziyuan Yang, Yingyu Chen, Zhiwen Wang, Hongming Shan, Yang Chen, Yi
Zhang
|
Patient-Level Anatomy Meets Scanning-Level Physics: Personalized
Federated Low-Dose CT Denoising Empowered by Large Language Model
|
Accepted by CVPR 2025
| null | null | null |
eess.IV cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Reducing radiation doses benefits patients, however, the resultant low-dose
computed tomography (LDCT) images often suffer from clinically unacceptable
noise and artifacts. While deep learning (DL) shows promise in LDCT
reconstruction, it requires large-scale data collection from multiple clients,
raising privacy concerns. Federated learning (FL) has been introduced to
address these privacy concerns; however, current methods are typically tailored
to specific scanning protocols, which limits their generalizability and makes
them less effective for unseen protocols. To address these issues, we propose
SCAN-PhysFed, a novel SCanning- and ANatomy-level personalized Physics-Driven
Federated learning paradigm for LDCT reconstruction. Since the noise
distribution in LDCT data is closely tied to scanning protocols and anatomical
structures being scanned, we design a dual-level physics-informed way to
address these challenges. Specifically, we incorporate physical and anatomical
prompts into our physics-informed hypernetworks to capture scanning- and
anatomy-specific information, enabling dual-level physics-driven
personalization of imaging features. These prompts are derived from the
scanning protocol and the radiology report generated by a medical large
language model (MLLM), respectively. Subsequently, client-specific decoders
project these dual-level personalized imaging features back into the image
domain. Besides, to tackle the challenge of unseen data, we introduce a novel
protocol vector-quantization strategy (PVQS), which ensures consistent
performance across new clients by quantifying the unseen scanning code as one
of the codes in the scanning codebook. Extensive experimental results
demonstrate the superior performance of SCAN-PhysFed on public datasets.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 14:20:32 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Yang",
"Ziyuan",
""
],
[
"Chen",
"Yingyu",
""
],
[
"Wang",
"Zhiwen",
""
],
[
"Shan",
"Hongming",
""
],
[
"Chen",
"Yang",
""
],
[
"Zhang",
"Yi",
""
]
] |
TITLE: Patient-Level Anatomy Meets Scanning-Level Physics: Personalized
Federated Low-Dose CT Denoising Empowered by Large Language Model
ABSTRACT: Reducing radiation doses benefits patients, however, the resultant low-dose
computed tomography (LDCT) images often suffer from clinically unacceptable
noise and artifacts. While deep learning (DL) shows promise in LDCT
reconstruction, it requires large-scale data collection from multiple clients,
raising privacy concerns. Federated learning (FL) has been introduced to
address these privacy concerns; however, current methods are typically tailored
to specific scanning protocols, which limits their generalizability and makes
them less effective for unseen protocols. To address these issues, we propose
SCAN-PhysFed, a novel SCanning- and ANatomy-level personalized Physics-Driven
Federated learning paradigm for LDCT reconstruction. Since the noise
distribution in LDCT data is closely tied to scanning protocols and anatomical
structures being scanned, we design a dual-level physics-informed way to
address these challenges. Specifically, we incorporate physical and anatomical
prompts into our physics-informed hypernetworks to capture scanning- and
anatomy-specific information, enabling dual-level physics-driven
personalization of imaging features. These prompts are derived from the
scanning protocol and the radiology report generated by a medical large
language model (MLLM), respectively. Subsequently, client-specific decoders
project these dual-level personalized imaging features back into the image
domain. Besides, to tackle the challenge of unseen data, we introduce a novel
protocol vector-quantization strategy (PVQS), which ensures consistent
performance across new clients by quantifying the unseen scanning code as one
of the codes in the scanning codebook. Extensive experimental results
demonstrate the superior performance of SCAN-PhysFed on public datasets.
|
no_new_dataset
| 0.957991 |
2503.00912
|
Zhuohang Jiang
|
Zhuohang Jiang, Pangjing Wu, Ziran Liang, Peter Q. Chen, Xu Yuan, Ye
Jia, Jiancheng Tu, Chen Li, Peter H.F. Ng, Qing Li
|
HiBench: Benchmarking LLMs Capability on Hierarchical Structure
Reasoning
| null | null | null | null |
cs.CL cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Structure reasoning is a fundamental capability of large language models
(LLMs), enabling them to reason about structured commonsense and answer
multi-hop questions. However, existing benchmarks for structure reasoning
mainly focus on horizontal and coordinate structures (\emph{e.g.} graphs),
overlooking the hierarchical relationships within them. Hierarchical structure
reasoning is crucial for human cognition, particularly in memory organization
and problem-solving. It also plays a key role in various real-world tasks, such
as information extraction and decision-making. To address this gap, we propose
HiBench, the first framework spanning from initial structure generation to
final proficiency assessment, designed to benchmark the hierarchical reasoning
capabilities of LLMs systematically. HiBench encompasses six representative
scenarios, covering both fundamental and practical aspects, and consists of 30
tasks with varying hierarchical complexity, totaling 39,519 queries. To
evaluate LLMs comprehensively, we develop five capability dimensions that
depict different facets of hierarchical structure understanding. Through
extensive evaluation of 20 LLMs from 10 model families, we reveal key insights
into their capabilities and limitations: 1) existing LLMs show proficiency in
basic hierarchical reasoning tasks; 2) they still struggle with more complex
structures and implicit hierarchical representations, especially in structural
modification and textual reasoning. Based on these findings, we create a small
yet well-designed instruction dataset, which enhances LLMs' performance on
HiBench by an average of 88.84\% (Llama-3.1-8B) and 31.38\% (Qwen2.5-7B) across
all tasks. The HiBench dataset and toolkit are available here,
https://github.com/jzzzzh/HiBench, to encourage evaluation.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 14:25:37 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Jiang",
"Zhuohang",
""
],
[
"Wu",
"Pangjing",
""
],
[
"Liang",
"Ziran",
""
],
[
"Chen",
"Peter Q.",
""
],
[
"Yuan",
"Xu",
""
],
[
"Jia",
"Ye",
""
],
[
"Tu",
"Jiancheng",
""
],
[
"Li",
"Chen",
""
],
[
"Ng",
"Peter H. F.",
""
],
[
"Li",
"Qing",
""
]
] |
TITLE: HiBench: Benchmarking LLMs Capability on Hierarchical Structure
Reasoning
ABSTRACT: Structure reasoning is a fundamental capability of large language models
(LLMs), enabling them to reason about structured commonsense and answer
multi-hop questions. However, existing benchmarks for structure reasoning
mainly focus on horizontal and coordinate structures (\emph{e.g.} graphs),
overlooking the hierarchical relationships within them. Hierarchical structure
reasoning is crucial for human cognition, particularly in memory organization
and problem-solving. It also plays a key role in various real-world tasks, such
as information extraction and decision-making. To address this gap, we propose
HiBench, the first framework spanning from initial structure generation to
final proficiency assessment, designed to benchmark the hierarchical reasoning
capabilities of LLMs systematically. HiBench encompasses six representative
scenarios, covering both fundamental and practical aspects, and consists of 30
tasks with varying hierarchical complexity, totaling 39,519 queries. To
evaluate LLMs comprehensively, we develop five capability dimensions that
depict different facets of hierarchical structure understanding. Through
extensive evaluation of 20 LLMs from 10 model families, we reveal key insights
into their capabilities and limitations: 1) existing LLMs show proficiency in
basic hierarchical reasoning tasks; 2) they still struggle with more complex
structures and implicit hierarchical representations, especially in structural
modification and textual reasoning. Based on these findings, we create a small
yet well-designed instruction dataset, which enhances LLMs' performance on
HiBench by an average of 88.84\% (Llama-3.1-8B) and 31.38\% (Qwen2.5-7B) across
all tasks. The HiBench dataset and toolkit are available here,
https://github.com/jzzzzh/HiBench, to encourage evaluation.
|
new_dataset
| 0.965053 |
2503.00915
|
Xitong Ling
|
Xitong Ling, Yifeng Ping, Jiawen Li, Jing Peng, Yuxuan Chen, Minxi
Ouyang, Yizhi Wang, Yonghong He, Tian Guan, Xiaoping Liu, Lianghui Zhu
|
Multimodal Distillation-Driven Ensemble Learning for Long-Tailed
Histopathology Whole Slide Images Analysis
| null | null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Multiple Instance Learning (MIL) plays a significant role in computational
pathology, enabling weakly supervised analysis of Whole Slide Image (WSI)
datasets. The field of WSI analysis is confronted with a severe long-tailed
distribution problem, which significantly impacts the performance of
classifiers. Long-tailed distributions lead to class imbalance, where some
classes have sparse samples while others are abundant, making it difficult for
classifiers to accurately identify minority class samples. To address this
issue, we propose an ensemble learning method based on MIL, which employs
expert decoders with shared aggregators and consistency constraints to learn
diverse distributions and reduce the impact of class imbalance on classifier
performance. Moreover, we introduce a multimodal distillation framework that
leverages text encoders pre-trained on pathology-text pairs to distill
knowledge and guide the MIL aggregator in capturing stronger semantic features
relevant to class information. To ensure flexibility, we use learnable prompts
to guide the distillation process of the pre-trained text encoder, avoiding
limitations imposed by specific prompts. Our method, MDE-MIL, integrates
multiple expert branches focusing on specific data distributions to address
long-tailed issues. Consistency control ensures generalization across classes.
Multimodal distillation enhances feature extraction. Experiments on
Camelyon+-LT and PANDA-LT datasets show it outperforms state-of-the-art
methods.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 14:31:45 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ling",
"Xitong",
""
],
[
"Ping",
"Yifeng",
""
],
[
"Li",
"Jiawen",
""
],
[
"Peng",
"Jing",
""
],
[
"Chen",
"Yuxuan",
""
],
[
"Ouyang",
"Minxi",
""
],
[
"Wang",
"Yizhi",
""
],
[
"He",
"Yonghong",
""
],
[
"Guan",
"Tian",
""
],
[
"Liu",
"Xiaoping",
""
],
[
"Zhu",
"Lianghui",
""
]
] |
TITLE: Multimodal Distillation-Driven Ensemble Learning for Long-Tailed
Histopathology Whole Slide Images Analysis
ABSTRACT: Multiple Instance Learning (MIL) plays a significant role in computational
pathology, enabling weakly supervised analysis of Whole Slide Image (WSI)
datasets. The field of WSI analysis is confronted with a severe long-tailed
distribution problem, which significantly impacts the performance of
classifiers. Long-tailed distributions lead to class imbalance, where some
classes have sparse samples while others are abundant, making it difficult for
classifiers to accurately identify minority class samples. To address this
issue, we propose an ensemble learning method based on MIL, which employs
expert decoders with shared aggregators and consistency constraints to learn
diverse distributions and reduce the impact of class imbalance on classifier
performance. Moreover, we introduce a multimodal distillation framework that
leverages text encoders pre-trained on pathology-text pairs to distill
knowledge and guide the MIL aggregator in capturing stronger semantic features
relevant to class information. To ensure flexibility, we use learnable prompts
to guide the distillation process of the pre-trained text encoder, avoiding
limitations imposed by specific prompts. Our method, MDE-MIL, integrates
multiple expert branches focusing on specific data distributions to address
long-tailed issues. Consistency control ensures generalization across classes.
Multimodal distillation enhances feature extraction. Experiments on
Camelyon+-LT and PANDA-LT datasets show it outperforms state-of-the-art
methods.
|
no_new_dataset
| 0.94625 |
2503.00917
|
Ali Ebrahimpour-Boroojeny
|
Ali Ebrahimpour-Boroojeny, Hari Sundaram, and Varun Chandrasekaran
|
AMUN: Adversarial Machine UNlearning
| null | null | null | null |
cs.LG cs.CR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Machine unlearning, where users can request the deletion of a forget dataset,
is becoming increasingly important because of numerous privacy regulations.
Initial works on ``exact'' unlearning (e.g., retraining) incur large
computational overheads. However, while computationally inexpensive,
``approximate'' methods have fallen short of reaching the effectiveness of
exact unlearning: models produced fail to obtain comparable accuracy and
prediction confidence on both the forget and test (i.e., unseen) dataset.
Exploiting this observation, we propose a new unlearning method, Adversarial
Machine UNlearning (AMUN), that outperforms prior state-of-the-art (SOTA)
methods for image classification. AMUN lowers the confidence of the model on
the forget samples by fine-tuning the model on their corresponding adversarial
examples. Adversarial examples naturally belong to the distribution imposed by
the model on the input space; fine-tuning the model on the adversarial examples
closest to the corresponding forget samples (a) localizes the changes to the
decision boundary of the model around each forget sample and (b) avoids drastic
changes to the global behavior of the model, thereby preserving the model's
accuracy on test samples. Using AMUN for unlearning a random $10\%$ of CIFAR-10
samples, we observe that even SOTA membership inference attacks cannot do
better than random guessing.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 14:36:31 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ebrahimpour-Boroojeny",
"Ali",
""
],
[
"Sundaram",
"Hari",
""
],
[
"Chandrasekaran",
"Varun",
""
]
] |
TITLE: AMUN: Adversarial Machine UNlearning
ABSTRACT: Machine unlearning, where users can request the deletion of a forget dataset,
is becoming increasingly important because of numerous privacy regulations.
Initial works on ``exact'' unlearning (e.g., retraining) incur large
computational overheads. However, while computationally inexpensive,
``approximate'' methods have fallen short of reaching the effectiveness of
exact unlearning: models produced fail to obtain comparable accuracy and
prediction confidence on both the forget and test (i.e., unseen) dataset.
Exploiting this observation, we propose a new unlearning method, Adversarial
Machine UNlearning (AMUN), that outperforms prior state-of-the-art (SOTA)
methods for image classification. AMUN lowers the confidence of the model on
the forget samples by fine-tuning the model on their corresponding adversarial
examples. Adversarial examples naturally belong to the distribution imposed by
the model on the input space; fine-tuning the model on the adversarial examples
closest to the corresponding forget samples (a) localizes the changes to the
decision boundary of the model around each forget sample and (b) avoids drastic
changes to the global behavior of the model, thereby preserving the model's
accuracy on test samples. Using AMUN for unlearning a random $10\%$ of CIFAR-10
samples, we observe that even SOTA membership inference attacks cannot do
better than random guessing.
|
no_new_dataset
| 0.948537 |
2503.00924
|
Daolang Huang
|
Xinyu Zhang, Daolang Huang, Samuel Kaski, Julien Martinelli
|
PABBO: Preferential Amortized Black-Box Optimization
|
25 pages, 17 figures. Accepted at the Thirteenth International
Conference on Learning Representations (ICLR 2025)
| null | null | null |
stat.ML cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Preferential Bayesian Optimization (PBO) is a sample-efficient method to
learn latent user utilities from preferential feedback over a pair of designs.
It relies on a statistical surrogate model for the latent function, usually a
Gaussian process, and an acquisition strategy to select the next candidate pair
to get user feedback on. Due to the non-conjugacy of the associated likelihood,
every PBO step requires a significant amount of computations with various
approximate inference techniques. This computational overhead is incompatible
with the way humans interact with computers, hindering the use of PBO in
real-world cases. Building on the recent advances of amortized BO, we propose
to circumvent this issue by fully amortizing PBO, meta-learning both the
surrogate and the acquisition function. Our method comprises a novel
transformer neural process architecture, trained using reinforcement learning
and tailored auxiliary losses. On a benchmark composed of synthetic and
real-world datasets, our method is several orders of magnitude faster than the
usual Gaussian process-based strategies and often outperforms them in accuracy.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 14:57:24 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhang",
"Xinyu",
""
],
[
"Huang",
"Daolang",
""
],
[
"Kaski",
"Samuel",
""
],
[
"Martinelli",
"Julien",
""
]
] |
TITLE: PABBO: Preferential Amortized Black-Box Optimization
ABSTRACT: Preferential Bayesian Optimization (PBO) is a sample-efficient method to
learn latent user utilities from preferential feedback over a pair of designs.
It relies on a statistical surrogate model for the latent function, usually a
Gaussian process, and an acquisition strategy to select the next candidate pair
to get user feedback on. Due to the non-conjugacy of the associated likelihood,
every PBO step requires a significant amount of computations with various
approximate inference techniques. This computational overhead is incompatible
with the way humans interact with computers, hindering the use of PBO in
real-world cases. Building on the recent advances of amortized BO, we propose
to circumvent this issue by fully amortizing PBO, meta-learning both the
surrogate and the acquisition function. Our method comprises a novel
transformer neural process architecture, trained using reinforcement learning
and tailored auxiliary losses. On a benchmark composed of synthetic and
real-world datasets, our method is several orders of magnitude faster than the
usual Gaussian process-based strategies and often outperforms them in accuracy.
|
no_new_dataset
| 0.946498 |
2503.00925
|
Daiki Nishiyama
|
Daiki Nishiyama, Hiroaki Miyoshi, Noriaki Hashimoto, Koichi Ohshima,
Hidekata Hontani, Ichiro Takeuchi, Jun Sakuma
|
Explainable Classifier for Malignant Lymphoma Subtyping via Cell Graph
and Image Fusion
|
11 pages, 3 figure
| null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Malignant lymphoma subtype classification directly impacts treatment
strategies and patient outcomes, necessitating classification models that
achieve both high accuracy and sufficient explainability. This study proposes a
novel explainable Multi-Instance Learning (MIL) framework that identifies
subtype-specific Regions of Interest (ROIs) from Whole Slide Images (WSIs)
while integrating cell distribution characteristics and image information. Our
framework simultaneously addresses three objectives: (1) indicating appropriate
ROIs for each subtype, (2) explaining the frequency and spatial distribution of
characteristic cell types, and (3) achieving high-accuracy subtyping by
leveraging both image and cell-distribution modalities. The proposed method
fuses cell graph and image features extracted from each patch in the WSI using
a Mixture-of-Experts (MoE) approach and classifies subtypes within an MIL
framework. Experiments on a dataset of 1,233 WSIs demonstrate that our approach
achieves state-of-the-art accuracy among ten comparative methods and provides
region-level and cell-level explanations that align with a pathologist's
perspectives.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 15:04:10 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Nishiyama",
"Daiki",
""
],
[
"Miyoshi",
"Hiroaki",
""
],
[
"Hashimoto",
"Noriaki",
""
],
[
"Ohshima",
"Koichi",
""
],
[
"Hontani",
"Hidekata",
""
],
[
"Takeuchi",
"Ichiro",
""
],
[
"Sakuma",
"Jun",
""
]
] |
TITLE: Explainable Classifier for Malignant Lymphoma Subtyping via Cell Graph
and Image Fusion
ABSTRACT: Malignant lymphoma subtype classification directly impacts treatment
strategies and patient outcomes, necessitating classification models that
achieve both high accuracy and sufficient explainability. This study proposes a
novel explainable Multi-Instance Learning (MIL) framework that identifies
subtype-specific Regions of Interest (ROIs) from Whole Slide Images (WSIs)
while integrating cell distribution characteristics and image information. Our
framework simultaneously addresses three objectives: (1) indicating appropriate
ROIs for each subtype, (2) explaining the frequency and spatial distribution of
characteristic cell types, and (3) achieving high-accuracy subtyping by
leveraging both image and cell-distribution modalities. The proposed method
fuses cell graph and image features extracted from each patch in the WSI using
a Mixture-of-Experts (MoE) approach and classifies subtypes within an MIL
framework. Experiments on a dataset of 1,233 WSIs demonstrate that our approach
achieves state-of-the-art accuracy among ten comparative methods and provides
region-level and cell-level explanations that align with a pathologist's
perspectives.
|
no_new_dataset
| 0.94743 |
2503.00930
|
Padmanaba Srinivasan
|
Padmanaba Srinivasan, William Knottenbelt
|
Behavior Preference Regression for Offline Reinforcement Learning
|
Conference paper at AAAI 25
| null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Offline reinforcement learning (RL) methods aim to learn optimal policies
with access only to trajectories in a fixed dataset. Policy constraint methods
formulate policy learning as an optimization problem that balances maximizing
reward with minimizing deviation from the behavior policy. Closed form
solutions to this problem can be derived as weighted behavioral cloning
objectives that, in theory, must compute an intractable partition function.
Reinforcement learning has gained popularity in language modeling to align
models with human preferences; some recent works consider paired completions
that are ranked by a preference model following which the likelihood of the
preferred completion is directly increased. We adapt this approach of paired
comparison. By reformulating the paired-sample optimization problem, we fit the
maximum-mode of the Q function while maximizing behavioral consistency of
policy actions. This yields our algorithm, Behavior Preference Regression for
offline RL (BPR). We empirically evaluate BPR on the widely used D4RL
Locomotion and Antmaze datasets, as well as the more challenging V-D4RL suite,
which operates in image-based state spaces. BPR demonstrates state-of-the-art
performance over all domains. Our on-policy experiments suggest that BPR takes
advantage of the stability of on-policy value functions with minimal
perceptible performance degradation on Locomotion datasets.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 15:13:02 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Srinivasan",
"Padmanaba",
""
],
[
"Knottenbelt",
"William",
""
]
] |
TITLE: Behavior Preference Regression for Offline Reinforcement Learning
ABSTRACT: Offline reinforcement learning (RL) methods aim to learn optimal policies
with access only to trajectories in a fixed dataset. Policy constraint methods
formulate policy learning as an optimization problem that balances maximizing
reward with minimizing deviation from the behavior policy. Closed form
solutions to this problem can be derived as weighted behavioral cloning
objectives that, in theory, must compute an intractable partition function.
Reinforcement learning has gained popularity in language modeling to align
models with human preferences; some recent works consider paired completions
that are ranked by a preference model following which the likelihood of the
preferred completion is directly increased. We adapt this approach of paired
comparison. By reformulating the paired-sample optimization problem, we fit the
maximum-mode of the Q function while maximizing behavioral consistency of
policy actions. This yields our algorithm, Behavior Preference Regression for
offline RL (BPR). We empirically evaluate BPR on the widely used D4RL
Locomotion and Antmaze datasets, as well as the more challenging V-D4RL suite,
which operates in image-based state spaces. BPR demonstrates state-of-the-art
performance over all domains. Our on-policy experiments suggest that BPR takes
advantage of the stability of on-policy value functions with minimal
perceptible performance degradation on Locomotion datasets.
|
no_new_dataset
| 0.941169 |
2503.00932
|
Qing Wan
|
Qing Wan, Shilong Deng, Xun Wang
|
Improving the Transferability of Adversarial Attacks by an Input
Transpose
|
15 pages, 11 figures
| null | null | null |
cs.CV cs.AI cs.CR cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Deep neural networks (DNNs) are highly susceptible to adversarial
examples--subtle perturbations applied to inputs that are often imperceptible
to humans yet lead to incorrect model predictions. In black-box scenarios,
however, existing adversarial examples exhibit limited transferability and
struggle to effectively compromise multiple unseen DNN models. Previous
strategies enhance the cross-model generalization of adversarial examples by
introducing versatility into adversarial perturbations, thereby improving
transferability. However, further refining perturbation versatility often
demands intricate algorithm development and substantial computation
consumption. In this work, we propose an input transpose method that requires
almost no additional labor and computation costs but can significantly improve
the transferability of existing adversarial strategies. Even without adding
adversarial perturbations, our method demonstrates considerable effectiveness
in cross-model attacks. Our exploration finds that on specific datasets, a mere
$1^\circ$ left or right rotation might be sufficient for most adversarial
examples to deceive unseen models. Our further analysis suggests that this
transferability improvement triggered by rotating only $1^\circ$ may stem from
visible pattern shifts in the DNN's low-level feature maps. Moreover, this
transferability exhibits optimal angles that, when identified under
unrestricted query conditions, could potentially yield even greater
performance.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 15:13:41 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Wan",
"Qing",
""
],
[
"Deng",
"Shilong",
""
],
[
"Wang",
"Xun",
""
]
] |
TITLE: Improving the Transferability of Adversarial Attacks by an Input
Transpose
ABSTRACT: Deep neural networks (DNNs) are highly susceptible to adversarial
examples--subtle perturbations applied to inputs that are often imperceptible
to humans yet lead to incorrect model predictions. In black-box scenarios,
however, existing adversarial examples exhibit limited transferability and
struggle to effectively compromise multiple unseen DNN models. Previous
strategies enhance the cross-model generalization of adversarial examples by
introducing versatility into adversarial perturbations, thereby improving
transferability. However, further refining perturbation versatility often
demands intricate algorithm development and substantial computation
consumption. In this work, we propose an input transpose method that requires
almost no additional labor and computation costs but can significantly improve
the transferability of existing adversarial strategies. Even without adding
adversarial perturbations, our method demonstrates considerable effectiveness
in cross-model attacks. Our exploration finds that on specific datasets, a mere
$1^\circ$ left or right rotation might be sufficient for most adversarial
examples to deceive unseen models. Our further analysis suggests that this
transferability improvement triggered by rotating only $1^\circ$ may stem from
visible pattern shifts in the DNN's low-level feature maps. Moreover, this
transferability exhibits optimal angles that, when identified under
unrestricted query conditions, could potentially yield even greater
performance.
|
no_new_dataset
| 0.940517 |
2503.00945
|
Hazrat Ali
|
Muhammad Rafiq, Hazrat Ali, Ghulam Mujtaba, Zubair Shah, Shoaib Azmat
|
Cross Modality Medical Image Synthesis for Improving Liver Segmentation
|
Submitted to Computer Methods in Biomechanics and Biomedical
Engineering: Imaging & Visualization
| null | null | null |
eess.IV cs.AI cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Deep learning-based computer-aided diagnosis (CAD) of medical images requires
large datasets. However, the lack of large publicly available labeled datasets
limits the development of deep learning-based CAD systems. Generative
Adversarial Networks (GANs), in particular, CycleGAN, can be used to generate
new cross-domain images without paired training data. However, most
CycleGAN-based synthesis methods lack the potential to overcome alignment and
asymmetry between the input and generated data. We propose a two-stage
technique for the synthesis of abdominal MRI using cross-modality translation
of abdominal CT. We show that the synthetic data can help improve the
performance of the liver segmentation network. We increase the number of
abdominal MRI images through cross-modality image transformation of unpaired CT
images using a CycleGAN inspired deformation invariant network called EssNet.
Subsequently, we combine the synthetic MRI images with the original MRI images
and use them to improve the accuracy of the U-Net on a liver segmentation task.
We train the U-Net on real MRI images and then on real and synthetic MRI
images. Consequently, by comparing both scenarios, we achieve an improvement in
the performance of U-Net. In summary, the improvement achieved in the
Intersection over Union (IoU) is 1.17%. The results show potential to address
the data scarcity challenge in medical imaging.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 15:54:12 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Rafiq",
"Muhammad",
""
],
[
"Ali",
"Hazrat",
""
],
[
"Mujtaba",
"Ghulam",
""
],
[
"Shah",
"Zubair",
""
],
[
"Azmat",
"Shoaib",
""
]
] |
TITLE: Cross Modality Medical Image Synthesis for Improving Liver Segmentation
ABSTRACT: Deep learning-based computer-aided diagnosis (CAD) of medical images requires
large datasets. However, the lack of large publicly available labeled datasets
limits the development of deep learning-based CAD systems. Generative
Adversarial Networks (GANs), in particular, CycleGAN, can be used to generate
new cross-domain images without paired training data. However, most
CycleGAN-based synthesis methods lack the potential to overcome alignment and
asymmetry between the input and generated data. We propose a two-stage
technique for the synthesis of abdominal MRI using cross-modality translation
of abdominal CT. We show that the synthetic data can help improve the
performance of the liver segmentation network. We increase the number of
abdominal MRI images through cross-modality image transformation of unpaired CT
images using a CycleGAN inspired deformation invariant network called EssNet.
Subsequently, we combine the synthetic MRI images with the original MRI images
and use them to improve the accuracy of the U-Net on a liver segmentation task.
We train the U-Net on real MRI images and then on real and synthetic MRI
images. Consequently, by comparing both scenarios, we achieve an improvement in
the performance of U-Net. In summary, the improvement achieved in the
Intersection over Union (IoU) is 1.17%. The results show potential to address
the data scarcity challenge in medical imaging.
|
no_new_dataset
| 0.952042 |
2503.00958
|
Milad Alshomary Dr.
|
Milad Alshomary, Nikhil Reddy Varimalla, Vishal Anand and Kathleen
McKeown
|
Layered Insights: Generalizable Analysis of Authorial Style by
Leveraging All Transformer Layers
| null | null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
We propose a new approach for the authorship attribution task that leverages
the various linguistic representations learned at different layers of
pre-trained transformer-based models. We evaluate our approach on three
datasets, comparing it to a state-of-the-art baseline in in-domain and
out-of-domain scenarios. We found that utilizing various transformer layers
improves the robustness of authorship attribution models when tested on
out-of-domain data, resulting in new state-of-the-art results. Our analysis
gives further insights into how our model's different layers get specialized in
representing certain stylistic features that benefit the model when tested out
of the domain.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 16:47:31 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Alshomary",
"Milad",
""
],
[
"Varimalla",
"Nikhil Reddy",
""
],
[
"Anand",
"Vishal",
""
],
[
"McKeown",
"Kathleen",
""
]
] |
TITLE: Layered Insights: Generalizable Analysis of Authorial Style by
Leveraging All Transformer Layers
ABSTRACT: We propose a new approach for the authorship attribution task that leverages
the various linguistic representations learned at different layers of
pre-trained transformer-based models. We evaluate our approach on three
datasets, comparing it to a state-of-the-art baseline in in-domain and
out-of-domain scenarios. We found that utilizing various transformer layers
improves the robustness of authorship attribution models when tested on
out-of-domain data, resulting in new state-of-the-art results. Our analysis
gives further insights into how our model's different layers get specialized in
representing certain stylistic features that benefit the model when tested out
of the domain.
|
no_new_dataset
| 0.946745 |
2503.00962
|
Minh Vu
|
Minh H. Vu and Lorenzo Tronchin and Tufve Nyholm and Tommy L\"ofstedt
|
Using Synthetic Images to Augment Small Medical Image Datasets
|
14 pages
| null | null | null |
cs.CV cs.AI cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Recent years have witnessed a growing academic and industrial interest in
deep learning (DL) for medical imaging. To perform well, DL models require very
large labeled datasets. However, most medical imaging datasets are small, with
a limited number of annotated samples. The reason they are small is usually
because delineating medical images is time-consuming and demanding for
oncologists. There are various techniques that can be used to augment a
dataset, for example, to apply affine transformations or elastic
transformations to available images, or to add synthetic images generated by a
Generative Adversarial Network (GAN). In this work, we have developed a novel
conditional variant of a current GAN method, the StyleGAN2, to generate
multi-modal high-resolution medical images with the purpose to augment small
medical imaging datasets with these synthetic images. We use the synthetic and
real images from six datasets to train models for the downstream task of
semantic segmentation. The quality of the generated medical images and the
effect of this augmentation on the segmentation performance were evaluated
afterward. Finally, the results indicate that the downstream segmentation
models did not benefit from the generated images. Further work and analyses are
required to establish how this augmentation affects the segmentation
performance.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 17:02:11 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Vu",
"Minh H.",
""
],
[
"Tronchin",
"Lorenzo",
""
],
[
"Nyholm",
"Tufve",
""
],
[
"Löfstedt",
"Tommy",
""
]
] |
TITLE: Using Synthetic Images to Augment Small Medical Image Datasets
ABSTRACT: Recent years have witnessed a growing academic and industrial interest in
deep learning (DL) for medical imaging. To perform well, DL models require very
large labeled datasets. However, most medical imaging datasets are small, with
a limited number of annotated samples. The reason they are small is usually
because delineating medical images is time-consuming and demanding for
oncologists. There are various techniques that can be used to augment a
dataset, for example, to apply affine transformations or elastic
transformations to available images, or to add synthetic images generated by a
Generative Adversarial Network (GAN). In this work, we have developed a novel
conditional variant of a current GAN method, the StyleGAN2, to generate
multi-modal high-resolution medical images with the purpose to augment small
medical imaging datasets with these synthetic images. We use the synthetic and
real images from six datasets to train models for the downstream task of
semantic segmentation. The quality of the generated medical images and the
effect of this augmentation on the segmentation performance were evaluated
afterward. Finally, the results indicate that the downstream segmentation
models did not benefit from the generated images. Further work and analyses are
required to establish how this augmentation affects the segmentation
performance.
|
no_new_dataset
| 0.94887 |
2503.00971
|
Xiaohan Li
|
Xiaohan Li, Sebastian Pattinson
|
An Efficient and Uncertainty-aware Reinforcement Learning Framework for
Quality Assurance in Extrusion Additive Manufacturing
| null | null | null | null |
eess.SY cs.SY
|
http://creativecommons.org/licenses/by/4.0/
|
Defects in extrusion additive manufacturing remain common despite its
prevalent use. While numerous AI-driven approaches have been proposed to
improve quality assurance, the inherently dynamic nature of the printing
process poses persistent challenges. Regardless of how comprehensive the
training dataset is, out-of-distribution data remains inevitable. Consequently,
deterministic models often struggle to maintain robustness and, in some cases,
fail entirely when deployed in new or slightly altered printing environments.
This work introduces an agent that dynamically adjusts flow rate and
temperature setpoints in real time, optimizing process control while addressing
bottlenecks in training efficiency and uncertainty management. It integrates a
vision-based uncertainty quantification module with a reinforcement learning
controller, using probabilistic distributions to describe printing segments.
While the underlying networks are deterministic, these evolving distributions
introduce adaptability into the decision-making process. The vision system
classifies material extrusion with a certain level of precision, generating
corresponding distributions. A deep Q-learning controller interacts with a
simulated environment calibrated to the vision system precision, allowing the
agent to learn optimal actions while demonstrating appropriate hesitation when
necessary. By executing asynchronous actions and applying progressively
tightened elliptical reward shaping, the controller develops robust, adaptive
control strategies that account for the coupling effects between process
parameters. When deployed with zero-shot learning, the agent effectively
bridges the sim-to-real gap, correcting mild and severe under- and
over-extrusion reliably. Beyond extrusion additive manufacturing, this scalable
framework enables practical AI-driven quality assurance across various additive
manufacturing processes.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 17:47:08 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Li",
"Xiaohan",
""
],
[
"Pattinson",
"Sebastian",
""
]
] |
TITLE: An Efficient and Uncertainty-aware Reinforcement Learning Framework for
Quality Assurance in Extrusion Additive Manufacturing
ABSTRACT: Defects in extrusion additive manufacturing remain common despite its
prevalent use. While numerous AI-driven approaches have been proposed to
improve quality assurance, the inherently dynamic nature of the printing
process poses persistent challenges. Regardless of how comprehensive the
training dataset is, out-of-distribution data remains inevitable. Consequently,
deterministic models often struggle to maintain robustness and, in some cases,
fail entirely when deployed in new or slightly altered printing environments.
This work introduces an agent that dynamically adjusts flow rate and
temperature setpoints in real time, optimizing process control while addressing
bottlenecks in training efficiency and uncertainty management. It integrates a
vision-based uncertainty quantification module with a reinforcement learning
controller, using probabilistic distributions to describe printing segments.
While the underlying networks are deterministic, these evolving distributions
introduce adaptability into the decision-making process. The vision system
classifies material extrusion with a certain level of precision, generating
corresponding distributions. A deep Q-learning controller interacts with a
simulated environment calibrated to the vision system precision, allowing the
agent to learn optimal actions while demonstrating appropriate hesitation when
necessary. By executing asynchronous actions and applying progressively
tightened elliptical reward shaping, the controller develops robust, adaptive
control strategies that account for the coupling effects between process
parameters. When deployed with zero-shot learning, the agent effectively
bridges the sim-to-real gap, correcting mild and severe under- and
over-extrusion reliably. Beyond extrusion additive manufacturing, this scalable
framework enables practical AI-driven quality assurance across various additive
manufacturing processes.
|
no_new_dataset
| 0.947769 |
2503.00972
|
Wanwen Chen
|
Wanwen Chen, Carson Studders, Jamie J.Y. Kwon, Emily H.T. Pang, Eitan
Prisman and Septimiu E. Salcudean
|
Semantic-ICP: Iterative Closest Point for Non-rigid Multi-Organ Point
Cloud Registration
|
10 pages, 3 figures, submitted to MICCAI 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Point cloud registration is important in computer-aided interventions (CAI).
While learning-based point cloud registration methods have been developed,
their clinical application is hampered by issues of generalizability and
explainability. Therefore, classical point cloud registration methods, such as
Iterative Closest Point (ICP), are still widely applied in CAI. ICP methods
fail to consider that: (1) the points have well-defined semantic meaning, in
that each point can be related to a specific anatomical label; (2) the
deformation needs to follow biomechanical energy constraints. In this paper, we
present a novel semantic ICP (sem-ICP) method that handles multiple point
labels and uses linear elastic energy regularization. We use semantic labels to
improve the robustness of the closest point matching and propose a new point
cloud deformation representation to apply explicit biomechanical energy
regularization. Our experiments on the Learn2reg abdominal MR-CT registration
dataset and a trans-oral robotic surgery ultrasound-CT registration dataset
show that our method improves the Hausdorff distance compared with other
state-of-the-art ICP-based registration methods. We also perform a sensitivity
study to show that our rigid initialization achieves better convergence with
different initializations and visible ratios.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 17:50:52 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Chen",
"Wanwen",
""
],
[
"Studders",
"Carson",
""
],
[
"Kwon",
"Jamie J. Y.",
""
],
[
"Pang",
"Emily H. T.",
""
],
[
"Prisman",
"Eitan",
""
],
[
"Salcudean",
"Septimiu E.",
""
]
] |
TITLE: Semantic-ICP: Iterative Closest Point for Non-rigid Multi-Organ Point
Cloud Registration
ABSTRACT: Point cloud registration is important in computer-aided interventions (CAI).
While learning-based point cloud registration methods have been developed,
their clinical application is hampered by issues of generalizability and
explainability. Therefore, classical point cloud registration methods, such as
Iterative Closest Point (ICP), are still widely applied in CAI. ICP methods
fail to consider that: (1) the points have well-defined semantic meaning, in
that each point can be related to a specific anatomical label; (2) the
deformation needs to follow biomechanical energy constraints. In this paper, we
present a novel semantic ICP (sem-ICP) method that handles multiple point
labels and uses linear elastic energy regularization. We use semantic labels to
improve the robustness of the closest point matching and propose a new point
cloud deformation representation to apply explicit biomechanical energy
regularization. Our experiments on the Learn2reg abdominal MR-CT registration
dataset and a trans-oral robotic surgery ultrasound-CT registration dataset
show that our method improves the Hausdorff distance compared with other
state-of-the-art ICP-based registration methods. We also perform a sensitivity
study to show that our rigid initialization achieves better convergence with
different initializations and visible ratios.
|
no_new_dataset
| 0.949435 |
2503.01009
|
Jinzhao Li
|
Jinzhao Li, Nan Jiang, Yexiang Xue
|
An Exact Solver for Satisfiability Modulo Counting with Probabilistic
Circuits
| null | null | null | null |
cs.AI cs.LO
|
http://creativecommons.org/licenses/by/4.0/
|
Satisfiability Modulo Counting (SMC) is a recently proposed general language
to reason about problems integrating statistical and symbolic artificial
intelligence. An SMC formula is an extended SAT formula in which the truth
values of a few Boolean variables are determined by probabilistic inference.
Existing approximate solvers optimize surrogate objectives, which lack formal
guarantees. Current exact solvers directly integrate SAT solvers and
probabilistic inference solvers resulting in slow performance because of many
back-and-forth invocations of both solvers. We propose KOCO-SMC, an integrated
exact SMC solver that efficiently tracks lower and upper bounds in the
probabilistic inference process. It enhances computational efficiency by
enabling early estimation of probabilistic inference using only partial
variable assignments, whereas existing methods require full variable
assignments. In the experiment, we compare KOCO-SMC with currently available
approximate and exact SMC solvers on large-scale datasets and real-world
applications. Our approach delivers high-quality solutions with high
efficiency.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 20:28:20 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Li",
"Jinzhao",
""
],
[
"Jiang",
"Nan",
""
],
[
"Xue",
"Yexiang",
""
]
] |
TITLE: An Exact Solver for Satisfiability Modulo Counting with Probabilistic
Circuits
ABSTRACT: Satisfiability Modulo Counting (SMC) is a recently proposed general language
to reason about problems integrating statistical and symbolic artificial
intelligence. An SMC formula is an extended SAT formula in which the truth
values of a few Boolean variables are determined by probabilistic inference.
Existing approximate solvers optimize surrogate objectives, which lack formal
guarantees. Current exact solvers directly integrate SAT solvers and
probabilistic inference solvers resulting in slow performance because of many
back-and-forth invocations of both solvers. We propose KOCO-SMC, an integrated
exact SMC solver that efficiently tracks lower and upper bounds in the
probabilistic inference process. It enhances computational efficiency by
enabling early estimation of probabilistic inference using only partial
variable assignments, whereas existing methods require full variable
assignments. In the experiment, we compare KOCO-SMC with currently available
approximate and exact SMC solvers on large-scale datasets and real-world
applications. Our approach delivers high-quality solutions with high
efficiency.
|
no_new_dataset
| 0.940134 |
2503.01013
|
Yushan Jiang
|
Yushan Jiang, Wenchao Yu, Geon Lee, Dongjin Song, Kijung Shin, Wei
Cheng, Yanchi Liu, Haifeng Chen
|
Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Time series analysis provides essential insights for real-world system
dynamics and informs downstream decision-making, yet most existing methods
often overlook the rich contextual signals present in auxiliary modalities. To
bridge this gap, we introduce TimeXL, a multi-modal prediction framework that
integrates a prototype-based time series encoder with three collaborating Large
Language Models (LLMs) to deliver more accurate predictions and interpretable
explanations. First, a multi-modal prototype-based encoder processes both time
series and textual inputs to generate preliminary forecasts alongside
case-based rationales. These outputs then feed into a prediction LLM, which
refines the forecasts by reasoning over the encoder's predictions and
explanations. Next, a reflection LLM compares the predicted values against the
ground truth, identifying textual inconsistencies or noise. Guided by this
feedback, a refinement LLM iteratively enhances text quality and triggers
encoder retraining. This closed-loop workflow -- prediction, critique
(reflect), and refinement -- continuously boosts the framework's performance
and interpretability. Empirical evaluations on four real-world datasets
demonstrate that TimeXL achieves up to 8.9\% improvement in AUC and produces
human-centric, multi-modal explanations, highlighting the power of LLM-driven
reasoning for time series prediction.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 20:40:53 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Jiang",
"Yushan",
""
],
[
"Yu",
"Wenchao",
""
],
[
"Lee",
"Geon",
""
],
[
"Song",
"Dongjin",
""
],
[
"Shin",
"Kijung",
""
],
[
"Cheng",
"Wei",
""
],
[
"Liu",
"Yanchi",
""
],
[
"Chen",
"Haifeng",
""
]
] |
TITLE: Explainable Multi-modal Time Series Prediction with LLM-in-the-Loop
ABSTRACT: Time series analysis provides essential insights for real-world system
dynamics and informs downstream decision-making, yet most existing methods
often overlook the rich contextual signals present in auxiliary modalities. To
bridge this gap, we introduce TimeXL, a multi-modal prediction framework that
integrates a prototype-based time series encoder with three collaborating Large
Language Models (LLMs) to deliver more accurate predictions and interpretable
explanations. First, a multi-modal prototype-based encoder processes both time
series and textual inputs to generate preliminary forecasts alongside
case-based rationales. These outputs then feed into a prediction LLM, which
refines the forecasts by reasoning over the encoder's predictions and
explanations. Next, a reflection LLM compares the predicted values against the
ground truth, identifying textual inconsistencies or noise. Guided by this
feedback, a refinement LLM iteratively enhances text quality and triggers
encoder retraining. This closed-loop workflow -- prediction, critique
(reflect), and refinement -- continuously boosts the framework's performance
and interpretability. Empirical evaluations on four real-world datasets
demonstrate that TimeXL achieves up to 8.9\% improvement in AUC and produces
human-centric, multi-modal explanations, highlighting the power of LLM-driven
reasoning for time series prediction.
|
no_new_dataset
| 0.946597 |
2503.01022
|
Onur Boyar
|
Onur Boyar, Indra Priyadarsini, Seiji Takeda, Lisa Hamada
|
LLM-Fusion: A Novel Multimodal Fusion Model for Accelerated Material
Discovery
|
4 pages, presented at AAAI 2025 Workshop on AI to Accelerating
Science and Engineering (AI2ASE)
| null | null | null |
cond-mat.mtrl-sci cs.AI cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Discovering materials with desirable properties in an efficient way remains a
significant problem in materials science. Many studies have tackled this
problem by using different sets of information available about the materials.
Among them, multimodal approaches have been found to be promising because of
their ability to combine different sources of information. However, fusion
algorithms to date remain simple, lacking a mechanism to provide a rich
representation of multiple modalities. This paper presents LLM-Fusion, a novel
multimodal fusion model that leverages large language models (LLMs) to
integrate diverse representations, such as SMILES, SELFIES, text descriptions,
and molecular fingerprints, for accurate property prediction. Our approach
introduces a flexible LLM-based architecture that supports multimodal input
processing and enables material property prediction with higher accuracy than
traditional methods. We validate our model on two datasets across five
prediction tasks and demonstrate its effectiveness compared to unimodal and
naive concatenation baselines.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 21:13:04 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Boyar",
"Onur",
""
],
[
"Priyadarsini",
"Indra",
""
],
[
"Takeda",
"Seiji",
""
],
[
"Hamada",
"Lisa",
""
]
] |
TITLE: LLM-Fusion: A Novel Multimodal Fusion Model for Accelerated Material
Discovery
ABSTRACT: Discovering materials with desirable properties in an efficient way remains a
significant problem in materials science. Many studies have tackled this
problem by using different sets of information available about the materials.
Among them, multimodal approaches have been found to be promising because of
their ability to combine different sources of information. However, fusion
algorithms to date remain simple, lacking a mechanism to provide a rich
representation of multiple modalities. This paper presents LLM-Fusion, a novel
multimodal fusion model that leverages large language models (LLMs) to
integrate diverse representations, such as SMILES, SELFIES, text descriptions,
and molecular fingerprints, for accurate property prediction. Our approach
introduces a flexible LLM-based architecture that supports multimodal input
processing and enables material property prediction with higher accuracy than
traditional methods. We validate our model on two datasets across five
prediction tasks and demonstrate its effectiveness compared to unimodal and
naive concatenation baselines.
|
no_new_dataset
| 0.945147 |
2503.01046
|
Jiaqi Gu
|
Pingchuan Ma, Zhengqi Gao, Meng Zhang, Haoyu Yang, Mark Ren, Rena
Huang, Duane S. Boning, Jiaqi Gu
|
MAPS: Multi-Fidelity AI-Augmented Photonic Simulation and Inverse Design
Infrastructure
|
6 pages. Accepted to DATE 2025
| null | null | null |
physics.optics cs.AI cs.ET
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Inverse design has emerged as a transformative approach for photonic device
optimization, enabling the exploration of high-dimensional, non-intuitive
design spaces to create ultra-compact devices and advance photonic integrated
circuits (PICs) in computing and interconnects. However, practical challenges,
such as suboptimal device performance, limited manufacturability, high
sensitivity to variations, computational inefficiency, and lack of
interpretability, have hindered its adoption in commercial hardware. Recent
advancements in AI-assisted photonic simulation and design offer transformative
potential, accelerating simulations and design generation by orders of
magnitude over traditional numerical methods. Despite these breakthroughs, the
lack of an open-source, standardized infrastructure and evaluation benchmark
limits accessibility and cross-disciplinary collaboration. To address this, we
introduce MAPS, a multi-fidelity AI-augmented photonic simulation and inverse
design infrastructure designed to bridge this gap. MAPS features three
synergistic components: (1) MAPS-Data: A dataset acquisition framework for
generating multi-fidelity, richly labeled devices, providing high-quality data
for AI-for-optics research. (2) MAPS-Train: A flexible AI-for-photonics
training framework offering a hierarchical data loading pipeline, customizable
model construction, support for data- and physics-driven losses, and
comprehensive evaluations. (3) MAPS-InvDes: An advanced adjoint inverse design
toolkit that abstracts complex physics but exposes flexible optimization steps,
integrates pre-trained AI models, and incorporates fabrication variation
models. This infrastructure MAPS provides a unified, open-source platform for
developing, benchmarking, and advancing AI-assisted photonic design workflows,
accelerating innovation in photonic hardware optimization and scientific
machine learning.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 22:30:18 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ma",
"Pingchuan",
""
],
[
"Gao",
"Zhengqi",
""
],
[
"Zhang",
"Meng",
""
],
[
"Yang",
"Haoyu",
""
],
[
"Ren",
"Mark",
""
],
[
"Huang",
"Rena",
""
],
[
"Boning",
"Duane S.",
""
],
[
"Gu",
"Jiaqi",
""
]
] |
TITLE: MAPS: Multi-Fidelity AI-Augmented Photonic Simulation and Inverse Design
Infrastructure
ABSTRACT: Inverse design has emerged as a transformative approach for photonic device
optimization, enabling the exploration of high-dimensional, non-intuitive
design spaces to create ultra-compact devices and advance photonic integrated
circuits (PICs) in computing and interconnects. However, practical challenges,
such as suboptimal device performance, limited manufacturability, high
sensitivity to variations, computational inefficiency, and lack of
interpretability, have hindered its adoption in commercial hardware. Recent
advancements in AI-assisted photonic simulation and design offer transformative
potential, accelerating simulations and design generation by orders of
magnitude over traditional numerical methods. Despite these breakthroughs, the
lack of an open-source, standardized infrastructure and evaluation benchmark
limits accessibility and cross-disciplinary collaboration. To address this, we
introduce MAPS, a multi-fidelity AI-augmented photonic simulation and inverse
design infrastructure designed to bridge this gap. MAPS features three
synergistic components: (1) MAPS-Data: A dataset acquisition framework for
generating multi-fidelity, richly labeled devices, providing high-quality data
for AI-for-optics research. (2) MAPS-Train: A flexible AI-for-photonics
training framework offering a hierarchical data loading pipeline, customizable
model construction, support for data- and physics-driven losses, and
comprehensive evaluations. (3) MAPS-InvDes: An advanced adjoint inverse design
toolkit that abstracts complex physics but exposes flexible optimization steps,
integrates pre-trained AI models, and incorporates fabrication variation
models. This infrastructure MAPS provides a unified, open-source platform for
developing, benchmarking, and advancing AI-assisted photonic design workflows,
accelerating innovation in photonic hardware optimization and scientific
machine learning.
|
no_new_dataset
| 0.948537 |
2503.01057
|
Allen Paul
|
Allen Paul, Neill Campbell, Tony Shardlow
|
Sparse Randomized Approximation of Normal Cycles
| null | null | null | null |
math.NA cs.NA
|
http://creativecommons.org/licenses/by/4.0/
|
We develop a compression algorithm for the Normal-Cycles representations of
shape, using the Nystrom approximation in Reproducing Kernel Hilbert Spaces and
Ridge Leverage Score sampling. Our method has theoretical guarantees on the
rate of convergence of the compression error, and the obtained approximations
are shown to be useful for down-line tasks such as nonlinear shape registration
in the Large Deformation Metric Mapping (LDDMM) framework, even for very high
compression ratios. The performance of our algorithm is demonstrated on
large-scale shape data from modern geometry processing datasets, and is shown
to be fast and scalable with rapid error decay.
|
[
{
"version": "v1",
"created": "Sun, 2 Mar 2025 23:34:30 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Paul",
"Allen",
""
],
[
"Campbell",
"Neill",
""
],
[
"Shardlow",
"Tony",
""
]
] |
TITLE: Sparse Randomized Approximation of Normal Cycles
ABSTRACT: We develop a compression algorithm for the Normal-Cycles representations of
shape, using the Nystrom approximation in Reproducing Kernel Hilbert Spaces and
Ridge Leverage Score sampling. Our method has theoretical guarantees on the
rate of convergence of the compression error, and the obtained approximations
are shown to be useful for down-line tasks such as nonlinear shape registration
in the Large Deformation Metric Mapping (LDDMM) framework, even for very high
compression ratios. The performance of our algorithm is demonstrated on
large-scale shape data from modern geometry processing datasets, and is shown
to be fast and scalable with rapid error decay.
|
no_new_dataset
| 0.94699 |
2503.01067
|
Gokul Swamy
|
Gokul Swamy, Sanjiban Choudhury, Wen Sun, Zhiwei Steven Wu, J. Andrew
Bagnell
|
All Roads Lead to Likelihood: The Value of Reinforcement Learning in
Fine-Tuning
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by-sa/4.0/
|
From a first-principles perspective, it may seem odd that the strongest
results in foundation model fine-tuning (FT) are achieved via a relatively
complex, two-stage training procedure. Specifically, one first trains a reward
model (RM) on some dataset (e.g. human preferences) before using it to provide
online feedback as part of a downstream reinforcement learning (RL) procedure,
rather than directly optimizing the policy parameters on the dataset via
offline maximum likelihood estimation. In fact, from an information-theoretic
perspective, we can only lose information via passing through a reward model
and cannot create any new information via on-policy sampling. To explain this
discrepancy, we scrutinize several hypotheses on the value of RL in FT through
both theoretical and empirical lenses. Of the hypotheses considered, we find
the most support for the explanation that on problems with a
generation-verification gap, the combination of the ease of learning the
relatively simple RM (verifier) from the preference data, coupled with the
ability of the downstream RL procedure to then filter its search space to the
subset of policies (generators) that are optimal for relatively simple
verifiers is what leads to the superior performance of online FT.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 00:15:19 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Swamy",
"Gokul",
""
],
[
"Choudhury",
"Sanjiban",
""
],
[
"Sun",
"Wen",
""
],
[
"Wu",
"Zhiwei Steven",
""
],
[
"Bagnell",
"J. Andrew",
""
]
] |
TITLE: All Roads Lead to Likelihood: The Value of Reinforcement Learning in
Fine-Tuning
ABSTRACT: From a first-principles perspective, it may seem odd that the strongest
results in foundation model fine-tuning (FT) are achieved via a relatively
complex, two-stage training procedure. Specifically, one first trains a reward
model (RM) on some dataset (e.g. human preferences) before using it to provide
online feedback as part of a downstream reinforcement learning (RL) procedure,
rather than directly optimizing the policy parameters on the dataset via
offline maximum likelihood estimation. In fact, from an information-theoretic
perspective, we can only lose information via passing through a reward model
and cannot create any new information via on-policy sampling. To explain this
discrepancy, we scrutinize several hypotheses on the value of RL in FT through
both theoretical and empirical lenses. Of the hypotheses considered, we find
the most support for the explanation that on problems with a
generation-verification gap, the combination of the ease of learning the
relatively simple RM (verifier) from the preference data, coupled with the
ability of the downstream RL procedure to then filter its search space to the
subset of policies (generators) that are optimal for relatively simple
verifiers is what leads to the superior performance of online FT.
|
no_new_dataset
| 0.942771 |
2503.01072
|
Michael Smith
|
Yu Fu, Michael Stanley Smith and Anastasios Panagiotelis
|
Vector Copula Variational Inference and Dependent Block Posterior
Approximations
| null | null | null | null |
stat.ML cs.LG econ.EM stat.ME
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Variational inference (VI) is a popular method to estimate statistical and
econometric models. The key to VI is the selection of a tractable density to
approximate the Bayesian posterior. For large and complex models a common
choice is to assume independence between multivariate blocks in a partition of
the parameter space. While this simplifies the problem it can reduce accuracy.
This paper proposes using vector copulas to capture dependence between the
blocks parsimoniously. Tailored multivariate marginals are constructed using
learnable cyclically monotone transformations. We call the resulting joint
distribution a ``dependent block posterior'' approximation. Vector copula
models are suggested that make tractable and flexible variational
approximations. They allow for differing marginals, numbers of blocks, block
sizes and forms of between block dependence. They also allow for solution of
the variational optimization using fast and efficient stochastic gradient
methods. The efficacy and versatility of the approach is demonstrated using
four different statistical models and 16 datasets which have posteriors that
are challenging to approximate. In all cases, our method produces more accurate
posterior approximations than benchmark VI methods that either assume block
independence or factor-based dependence, at limited additional computational
cost.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 00:24:54 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Fu",
"Yu",
""
],
[
"Smith",
"Michael Stanley",
""
],
[
"Panagiotelis",
"Anastasios",
""
]
] |
TITLE: Vector Copula Variational Inference and Dependent Block Posterior
Approximations
ABSTRACT: Variational inference (VI) is a popular method to estimate statistical and
econometric models. The key to VI is the selection of a tractable density to
approximate the Bayesian posterior. For large and complex models a common
choice is to assume independence between multivariate blocks in a partition of
the parameter space. While this simplifies the problem it can reduce accuracy.
This paper proposes using vector copulas to capture dependence between the
blocks parsimoniously. Tailored multivariate marginals are constructed using
learnable cyclically monotone transformations. We call the resulting joint
distribution a ``dependent block posterior'' approximation. Vector copula
models are suggested that make tractable and flexible variational
approximations. They allow for differing marginals, numbers of blocks, block
sizes and forms of between block dependence. They also allow for solution of
the variational optimization using fast and efficient stochastic gradient
methods. The efficacy and versatility of the approach is demonstrated using
four different statistical models and 16 datasets which have posteriors that
are challenging to approximate. In all cases, our method produces more accurate
posterior approximations than benchmark VI methods that either assume block
independence or factor-based dependence, at limited additional computational
cost.
|
no_new_dataset
| 0.947866 |
2503.01075
|
Seunghoi Kim
|
Seunghoi Kim and Henry F. J. Tregidgo and Matteo Figini and Chen Jin
and Sarang Joshi and Daniel C. Alexander
|
Tackling Hallucination from Conditional Models for Medical Image
Reconstruction with DynamicDPS
| null | null | null | null |
eess.IV cs.AI cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Hallucinations are spurious structures not present in the ground truth,
posing a critical challenge in medical image reconstruction, especially for
data-driven conditional models. We hypothesize that combining an unconditional
diffusion model with data consistency, trained on a diverse dataset, can reduce
these hallucinations. Based on this, we propose DynamicDPS, a diffusion-based
framework that integrates conditional and unconditional diffusion models to
enhance low-quality medical images while systematically reducing
hallucinations. Our approach first generates an initial reconstruction using a
conditional model, then refines it with an adaptive diffusion-based inverse
problem solver. DynamicDPS skips early stage in the reverse process by
selecting an optimal starting time point per sample and applies Wolfe's line
search for adaptive step sizes, improving both efficiency and image fidelity.
Using diffusion priors and data consistency, our method effectively reduces
hallucinations from any conditional model output. We validate its effectiveness
in Image Quality Transfer for low-field MRI enhancement. Extensive evaluations
on synthetic and real MR scans, including a downstream task for tissue volume
estimation, show that DynamicDPS reduces hallucinations, improving relative
volume estimation by over 15% for critical tissues while using only 5% of the
sampling steps required by baseline diffusion models. As a model-agnostic and
fine-tuning-free approach, DynamicDPS offers a robust solution for
hallucination reduction in medical imaging. The code will be made publicly
available upon publication.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 00:33:04 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Kim",
"Seunghoi",
""
],
[
"Tregidgo",
"Henry F. J.",
""
],
[
"Figini",
"Matteo",
""
],
[
"Jin",
"Chen",
""
],
[
"Joshi",
"Sarang",
""
],
[
"Alexander",
"Daniel C.",
""
]
] |
TITLE: Tackling Hallucination from Conditional Models for Medical Image
Reconstruction with DynamicDPS
ABSTRACT: Hallucinations are spurious structures not present in the ground truth,
posing a critical challenge in medical image reconstruction, especially for
data-driven conditional models. We hypothesize that combining an unconditional
diffusion model with data consistency, trained on a diverse dataset, can reduce
these hallucinations. Based on this, we propose DynamicDPS, a diffusion-based
framework that integrates conditional and unconditional diffusion models to
enhance low-quality medical images while systematically reducing
hallucinations. Our approach first generates an initial reconstruction using a
conditional model, then refines it with an adaptive diffusion-based inverse
problem solver. DynamicDPS skips early stage in the reverse process by
selecting an optimal starting time point per sample and applies Wolfe's line
search for adaptive step sizes, improving both efficiency and image fidelity.
Using diffusion priors and data consistency, our method effectively reduces
hallucinations from any conditional model output. We validate its effectiveness
in Image Quality Transfer for low-field MRI enhancement. Extensive evaluations
on synthetic and real MR scans, including a downstream task for tissue volume
estimation, show that DynamicDPS reduces hallucinations, improving relative
volume estimation by over 15% for critical tissues while using only 5% of the
sampling steps required by baseline diffusion models. As a model-agnostic and
fine-tuning-free approach, DynamicDPS offers a robust solution for
hallucination reduction in medical imaging. The code will be made publicly
available upon publication.
|
no_new_dataset
| 0.950134 |
2503.01079
|
Asela Hevapathige
|
Asela Hevapathige, Ahad N. Zehmakan, Qing Wang
|
Depth-Adaptive Graph Neural Networks via Learnable Bakry-'Emery
Curvature
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Graph Neural Networks (GNNs) have demonstrated strong representation learning
capabilities for graph-based tasks. Recent advances on GNNs leverage geometric
properties, such as curvature, to enhance its representation capabilities by
modeling complex connectivity patterns and information flow within graphs.
However, most existing approaches focus solely on discrete graph topology,
overlooking diffusion dynamics and task-specific dependencies essential for
effective learning. To address this, we propose integrating Bakry-\'Emery
curvature, which captures both structural and task-driven aspects of
information propagation. We develop an efficient, learnable approximation
strategy, making curvature computation scalable for large graphs. Furthermore,
we introduce an adaptive depth mechanism that dynamically adjusts
message-passing layers per vertex based on its curvature, ensuring efficient
propagation. Our theoretical analysis establishes a link between curvature and
feature distinctiveness, showing that high-curvature vertices require fewer
layers, while low-curvature ones benefit from deeper propagation. Extensive
experiments on benchmark datasets validate the effectiveness of our approach,
showing consistent performance improvements across diverse graph learning
tasks.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 00:48:41 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Hevapathige",
"Asela",
""
],
[
"Zehmakan",
"Ahad N.",
""
],
[
"Wang",
"Qing",
""
]
] |
TITLE: Depth-Adaptive Graph Neural Networks via Learnable Bakry-'Emery
Curvature
ABSTRACT: Graph Neural Networks (GNNs) have demonstrated strong representation learning
capabilities for graph-based tasks. Recent advances on GNNs leverage geometric
properties, such as curvature, to enhance its representation capabilities by
modeling complex connectivity patterns and information flow within graphs.
However, most existing approaches focus solely on discrete graph topology,
overlooking diffusion dynamics and task-specific dependencies essential for
effective learning. To address this, we propose integrating Bakry-\'Emery
curvature, which captures both structural and task-driven aspects of
information propagation. We develop an efficient, learnable approximation
strategy, making curvature computation scalable for large graphs. Furthermore,
we introduce an adaptive depth mechanism that dynamically adjusts
message-passing layers per vertex based on its curvature, ensuring efficient
propagation. Our theoretical analysis establishes a link between curvature and
feature distinctiveness, showing that high-curvature vertices require fewer
layers, while low-curvature ones benefit from deeper propagation. Extensive
experiments on benchmark datasets validate the effectiveness of our approach,
showing consistent performance improvements across diverse graph learning
tasks.
|
no_new_dataset
| 0.947137 |
2503.01082
|
Yuchen Cao
|
Zhanyi Ding, Zhongyan Wang, Yeyubei Zhang, Yuchen Cao, Yunchong Liu,
Xiaorui Shen, Yexin Tian, Jianglai Dai
|
Efficient or Powerful? Trade-offs Between Machine Learning and Deep
Learning for Mental Illness Detection on Social Media
| null | null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
Social media platforms provide valuable insights into mental health trends by
capturing user-generated discussions on conditions such as depression, anxiety,
and suicidal ideation. Machine learning (ML) and deep learning (DL) models have
been increasingly applied to classify mental health conditions from textual
data, but selecting the most effective model involves trade-offs in accuracy,
interpretability, and computational efficiency. This study evaluates multiple
ML models, including logistic regression, random forest, and LightGBM,
alongside deep learning architectures such as ALBERT and Gated Recurrent Units
(GRUs), for both binary and multi-class classification of mental health
conditions. Our findings indicate that ML and DL models achieve comparable
classification performance on medium-sized datasets, with ML models offering
greater interpretability through variable importance scores, while DL models
are more robust to complex linguistic patterns. Additionally, ML models require
explicit feature engineering, whereas DL models learn hierarchical
representations directly from text. Logistic regression provides the advantage
of capturing both positive and negative associations between features and
mental health conditions, whereas tree-based models prioritize decision-making
power through split-based feature selection. This study offers empirical
insights into the advantages and limitations of different modeling approaches
and provides recommendations for selecting appropriate methods based on dataset
size, interpretability needs, and computational constraints.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 00:51:41 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ding",
"Zhanyi",
""
],
[
"Wang",
"Zhongyan",
""
],
[
"Zhang",
"Yeyubei",
""
],
[
"Cao",
"Yuchen",
""
],
[
"Liu",
"Yunchong",
""
],
[
"Shen",
"Xiaorui",
""
],
[
"Tian",
"Yexin",
""
],
[
"Dai",
"Jianglai",
""
]
] |
TITLE: Efficient or Powerful? Trade-offs Between Machine Learning and Deep
Learning for Mental Illness Detection on Social Media
ABSTRACT: Social media platforms provide valuable insights into mental health trends by
capturing user-generated discussions on conditions such as depression, anxiety,
and suicidal ideation. Machine learning (ML) and deep learning (DL) models have
been increasingly applied to classify mental health conditions from textual
data, but selecting the most effective model involves trade-offs in accuracy,
interpretability, and computational efficiency. This study evaluates multiple
ML models, including logistic regression, random forest, and LightGBM,
alongside deep learning architectures such as ALBERT and Gated Recurrent Units
(GRUs), for both binary and multi-class classification of mental health
conditions. Our findings indicate that ML and DL models achieve comparable
classification performance on medium-sized datasets, with ML models offering
greater interpretability through variable importance scores, while DL models
are more robust to complex linguistic patterns. Additionally, ML models require
explicit feature engineering, whereas DL models learn hierarchical
representations directly from text. Logistic regression provides the advantage
of capturing both positive and negative associations between features and
mental health conditions, whereas tree-based models prioritize decision-making
power through split-based feature selection. This study offers empirical
insights into the advantages and limitations of different modeling approaches
and provides recommendations for selecting appropriate methods based on dataset
size, interpretability needs, and computational constraints.
|
no_new_dataset
| 0.947624 |
2503.01085
|
Mykola Kozlenko
|
Mykola Kozlenko, Volodymyr Sendetskyi, Oleksiy Simkiv, Nazar
Savchenko, Andy Bosyi
|
Identity documents recognition and detection using semantic segmentation
with convolutional neural network
|
9 pages, 8 figures. This paper was originally published in 2021
Workshop on Cybersecurity Providing in Information and Telecommunication
Systems, in CEUR Workshop Proceedings, vol. 2923, available:
https://ceur-ws.org/Vol-2923/paper25.pdf
|
2021 Workshop on Cybersecurity Providing in Information and
Telecommunication Systems, in CEUR Workshop Proceedings, vol. 2923, Kyiv,
Ukraine, Jan. 28, 2021, pp. 234-242
| null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Object recognition and detection are well-studied problems with a developed
set of almost standard solutions. Identity documents recognition,
classification, detection, and localization are the tasks required in a number
of applications, particularly, in physical access control security systems at
critical infrastructure premises. In this paper, we propose the new original
architecture of a model based on an artificial convolutional neural network and
semantic segmentation approach for the recognition and detection of identity
documents in images. The challenge with the processing of such images is the
limited computational performance and the limited amount of memory when such an
application is running on industrial oneboard microcomputer hardware. The aim
of this research is to prove the feasibility of the proposed technique and to
obtain quality metrics. The methodology of the research is to evaluate the deep
learning detection model trained on the mobile identity document video dataset.
The dataset contains five hundred video clips for fifty different identity
document types. The numerical results from simulations are used to evaluate the
quality metrics. We present the results as accuracy versus threshold of the
intersection over union value. The paper reports an accuracy above 0.75 for the
intersection over union (IoU) threshold value of 0.8. Besides, we assessed the
size of the model and proved the feasibility of running the model on an
industrial one-board microcomputer or smartphone hardware.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 01:13:28 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Kozlenko",
"Mykola",
""
],
[
"Sendetskyi",
"Volodymyr",
""
],
[
"Simkiv",
"Oleksiy",
""
],
[
"Savchenko",
"Nazar",
""
],
[
"Bosyi",
"Andy",
""
]
] |
TITLE: Identity documents recognition and detection using semantic segmentation
with convolutional neural network
ABSTRACT: Object recognition and detection are well-studied problems with a developed
set of almost standard solutions. Identity documents recognition,
classification, detection, and localization are the tasks required in a number
of applications, particularly, in physical access control security systems at
critical infrastructure premises. In this paper, we propose the new original
architecture of a model based on an artificial convolutional neural network and
semantic segmentation approach for the recognition and detection of identity
documents in images. The challenge with the processing of such images is the
limited computational performance and the limited amount of memory when such an
application is running on industrial oneboard microcomputer hardware. The aim
of this research is to prove the feasibility of the proposed technique and to
obtain quality metrics. The methodology of the research is to evaluate the deep
learning detection model trained on the mobile identity document video dataset.
The dataset contains five hundred video clips for fifty different identity
document types. The numerical results from simulations are used to evaluate the
quality metrics. We present the results as accuracy versus threshold of the
intersection over union value. The paper reports an accuracy above 0.75 for the
intersection over union (IoU) threshold value of 0.8. Besides, we assessed the
size of the model and proved the feasibility of running the model on an
industrial one-board microcomputer or smartphone hardware.
|
new_dataset
| 0.972934 |
2503.01092
|
Kailun Yang
|
Wanjun Jia, Fan Yang, Mengfei Duan, Xianchi Chen, Yinxi Wang, Yiming
Jiang, Wenrui Chen, Kailun Yang, Zhiyong Li
|
One-Shot Affordance Grounding of Deformable Objects in Egocentric
Organizing Scenes
|
Source code and benchmark dataset will be publicly available at
https://github.com/Dikay1/OS-AGDO
| null | null | null |
cs.CV cs.RO eess.IV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Deformable object manipulation in robotics presents significant challenges
due to uncertainties in component properties, diverse configurations, visual
interference, and ambiguous prompts. These factors complicate both perception
and control tasks. To address these challenges, we propose a novel method for
One-Shot Affordance Grounding of Deformable Objects (OS-AGDO) in egocentric
organizing scenes, enabling robots to recognize previously unseen deformable
objects with varying colors and shapes using minimal samples. Specifically, we
first introduce the Deformable Object Semantic Enhancement Module (DefoSEM),
which enhances hierarchical understanding of the internal structure and
improves the ability to accurately identify local features, even under
conditions of weak component information. Next, we propose the ORB-Enhanced
Keypoint Fusion Module (OEKFM), which optimizes feature extraction of key
components by leveraging geometric constraints and improves adaptability to
diversity and visual interference. Additionally, we propose an
instance-conditional prompt based on image data and task context, effectively
mitigates the issue of region ambiguity caused by prompt words. To validate
these methods, we construct a diverse real-world dataset, AGDDO15, which
includes 15 common types of deformable objects and their associated
organizational actions. Experimental results demonstrate that our approach
significantly outperforms state-of-the-art methods, achieving improvements of
6.2%, 3.2%, and 2.9% in KLD, SIM, and NSS metrics, respectively, while
exhibiting high generalization performance. Source code and benchmark dataset
will be publicly available at https://github.com/Dikay1/OS-AGDO.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 01:34:56 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Jia",
"Wanjun",
""
],
[
"Yang",
"Fan",
""
],
[
"Duan",
"Mengfei",
""
],
[
"Chen",
"Xianchi",
""
],
[
"Wang",
"Yinxi",
""
],
[
"Jiang",
"Yiming",
""
],
[
"Chen",
"Wenrui",
""
],
[
"Yang",
"Kailun",
""
],
[
"Li",
"Zhiyong",
""
]
] |
TITLE: One-Shot Affordance Grounding of Deformable Objects in Egocentric
Organizing Scenes
ABSTRACT: Deformable object manipulation in robotics presents significant challenges
due to uncertainties in component properties, diverse configurations, visual
interference, and ambiguous prompts. These factors complicate both perception
and control tasks. To address these challenges, we propose a novel method for
One-Shot Affordance Grounding of Deformable Objects (OS-AGDO) in egocentric
organizing scenes, enabling robots to recognize previously unseen deformable
objects with varying colors and shapes using minimal samples. Specifically, we
first introduce the Deformable Object Semantic Enhancement Module (DefoSEM),
which enhances hierarchical understanding of the internal structure and
improves the ability to accurately identify local features, even under
conditions of weak component information. Next, we propose the ORB-Enhanced
Keypoint Fusion Module (OEKFM), which optimizes feature extraction of key
components by leveraging geometric constraints and improves adaptability to
diversity and visual interference. Additionally, we propose an
instance-conditional prompt based on image data and task context, effectively
mitigates the issue of region ambiguity caused by prompt words. To validate
these methods, we construct a diverse real-world dataset, AGDDO15, which
includes 15 common types of deformable objects and their associated
organizational actions. Experimental results demonstrate that our approach
significantly outperforms state-of-the-art methods, achieving improvements of
6.2%, 3.2%, and 2.9% in KLD, SIM, and NSS metrics, respectively, while
exhibiting high generalization performance. Source code and benchmark dataset
will be publicly available at https://github.com/Dikay1/OS-AGDO.
|
new_dataset
| 0.962321 |
2503.01097
|
Hyeon Jeon
|
Hyeon Jeon, Micha\"el Aupetit, DongHwa Shin, Aeri Cho, Seokhyeon Park,
Jinwook Seo
|
Measuring the Validity of Clustering Validation Datasets
|
IEEE Transactions on Pattern Analysis and Machine Intelligence
(TPAMI)
| null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Clustering techniques are often validated using benchmark datasets where
class labels are used as ground-truth clusters. However, depending on the
datasets, class labels may not align with the actual data clusters, and such
misalignment hampers accurate validation. Therefore, it is essential to
evaluate and compare datasets regarding their cluster-label matching (CLM),
i.e., how well their class labels match actual clusters. Internal validation
measures (IVMs), like Silhouette, can compare CLM over different labeling of
the same dataset, but are not designed to do so across different datasets. We
thus introduce Adjusted IVMs as fast and reliable methods to evaluate and
compare CLM across datasets. We establish four axioms that require validation
measures to be independent of data properties not related to cluster structure
(e.g., dimensionality, dataset size). Then, we develop standardized protocols
to convert any IVM to satisfy these axioms, and use these protocols to adjust
six widely used IVMs. Quantitative experiments (1) verify the necessity and
effectiveness of our protocols and (2) show that adjusted IVMs outperform the
competitors, including standard IVMs, in accurately evaluating CLM both within
and across datasets. We also show that the datasets can be filtered or improved
using our method to form more reliable benchmarks for clustering validation.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 01:54:04 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Jeon",
"Hyeon",
""
],
[
"Aupetit",
"Michaël",
""
],
[
"Shin",
"DongHwa",
""
],
[
"Cho",
"Aeri",
""
],
[
"Park",
"Seokhyeon",
""
],
[
"Seo",
"Jinwook",
""
]
] |
TITLE: Measuring the Validity of Clustering Validation Datasets
ABSTRACT: Clustering techniques are often validated using benchmark datasets where
class labels are used as ground-truth clusters. However, depending on the
datasets, class labels may not align with the actual data clusters, and such
misalignment hampers accurate validation. Therefore, it is essential to
evaluate and compare datasets regarding their cluster-label matching (CLM),
i.e., how well their class labels match actual clusters. Internal validation
measures (IVMs), like Silhouette, can compare CLM over different labeling of
the same dataset, but are not designed to do so across different datasets. We
thus introduce Adjusted IVMs as fast and reliable methods to evaluate and
compare CLM across datasets. We establish four axioms that require validation
measures to be independent of data properties not related to cluster structure
(e.g., dimensionality, dataset size). Then, we develop standardized protocols
to convert any IVM to satisfy these axioms, and use these protocols to adjust
six widely used IVMs. Quantitative experiments (1) verify the necessity and
effectiveness of our protocols and (2) show that adjusted IVMs outperform the
competitors, including standard IVMs, in accurately evaluating CLM both within
and across datasets. We also show that the datasets can be filtered or improved
using our method to form more reliable benchmarks for clustering validation.
|
no_new_dataset
| 0.944791 |
2503.01103
|
Kaiwen Zheng
|
Kaiwen Zheng, Yongxin Chen, Huayu Chen, Guande He, Ming-Yu Liu, Jun
Zhu, Qinsheng Zhang
|
Direct Discriminative Optimization: Your Likelihood-Based Visual
Generative Model is Secretly a GAN Discriminator
|
Project Page: https://research.nvidia.com/labs/dir/ddo/
| null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
While likelihood-based generative models, particularly diffusion and
autoregressive models, have achieved remarkable fidelity in visual generation,
the maximum likelihood estimation (MLE) objective inherently suffers from a
mode-covering tendency that limits the generation quality under limited model
capacity. In this work, we propose Direct Discriminative Optimization (DDO) as
a unified framework that bridges likelihood-based generative training and the
GAN objective to bypass this fundamental constraint. Our key insight is to
parameterize a discriminator implicitly using the likelihood ratio between a
learnable target model and a fixed reference model, drawing parallels with the
philosophy of Direct Preference Optimization (DPO). Unlike GANs, this
parameterization eliminates the need for joint training of generator and
discriminator networks, allowing for direct, efficient, and effective
finetuning of a well-trained model to its full potential beyond the limits of
MLE. DDO can be performed iteratively in a self-play manner for progressive
model refinement, with each round requiring less than 1% of pretraining epochs.
Our experiments demonstrate the effectiveness of DDO by significantly advancing
the previous SOTA diffusion model EDM, reducing FID scores from 1.79/1.58 to
new records of 1.30/0.97 on CIFAR-10/ImageNet-64 datasets, and by consistently
improving both guidance-free and CFG-enhanced FIDs of visual autoregressive
models on ImageNet 256$\times$256.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 02:06:22 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zheng",
"Kaiwen",
""
],
[
"Chen",
"Yongxin",
""
],
[
"Chen",
"Huayu",
""
],
[
"He",
"Guande",
""
],
[
"Liu",
"Ming-Yu",
""
],
[
"Zhu",
"Jun",
""
],
[
"Zhang",
"Qinsheng",
""
]
] |
TITLE: Direct Discriminative Optimization: Your Likelihood-Based Visual
Generative Model is Secretly a GAN Discriminator
ABSTRACT: While likelihood-based generative models, particularly diffusion and
autoregressive models, have achieved remarkable fidelity in visual generation,
the maximum likelihood estimation (MLE) objective inherently suffers from a
mode-covering tendency that limits the generation quality under limited model
capacity. In this work, we propose Direct Discriminative Optimization (DDO) as
a unified framework that bridges likelihood-based generative training and the
GAN objective to bypass this fundamental constraint. Our key insight is to
parameterize a discriminator implicitly using the likelihood ratio between a
learnable target model and a fixed reference model, drawing parallels with the
philosophy of Direct Preference Optimization (DPO). Unlike GANs, this
parameterization eliminates the need for joint training of generator and
discriminator networks, allowing for direct, efficient, and effective
finetuning of a well-trained model to its full potential beyond the limits of
MLE. DDO can be performed iteratively in a self-play manner for progressive
model refinement, with each round requiring less than 1% of pretraining epochs.
Our experiments demonstrate the effectiveness of DDO by significantly advancing
the previous SOTA diffusion model EDM, reducing FID scores from 1.79/1.58 to
new records of 1.30/0.97 on CIFAR-10/ImageNet-64 datasets, and by consistently
improving both guidance-free and CFG-enhanced FIDs of visual autoregressive
models on ImageNet 256$\times$256.
|
no_new_dataset
| 0.945851 |
2503.01109
|
Yansong Xu
|
Yansong Xu, Junlin Li, Wei Zhang, Siyu Chen, Shengyong Zhang, Yuquan
Leng, Weijia Zhou
|
FGS-SLAM: Fourier-based Gaussian Splatting for Real-time SLAM with
Sparse and Dense Map Fusion
| null | null | null | null |
cs.CV cs.AI cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
3D gaussian splatting has advanced simultaneous localization and mapping
(SLAM) technology by enabling real-time positioning and the construction of
high-fidelity maps. However, the uncertainty in gaussian position and
initialization parameters introduces challenges, often requiring extensive
iterative convergence and resulting in redundant or insufficient gaussian
representations. To address this, we introduce a novel adaptive densification
method based on Fourier frequency domain analysis to establish gaussian priors
for rapid convergence. Additionally, we propose constructing independent and
unified sparse and dense maps, where a sparse map supports efficient tracking
via Generalized Iterative Closest Point (GICP) and a dense map creates
high-fidelity visual representations. This is the first SLAM system leveraging
frequency domain analysis to achieve high-quality gaussian mapping in
real-time. Experimental results demonstrate an average frame rate of 36 FPS on
Replica and TUM RGB-D datasets, achieving competitive accuracy in both
localization and mapping.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 02:33:39 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Xu",
"Yansong",
""
],
[
"Li",
"Junlin",
""
],
[
"Zhang",
"Wei",
""
],
[
"Chen",
"Siyu",
""
],
[
"Zhang",
"Shengyong",
""
],
[
"Leng",
"Yuquan",
""
],
[
"Zhou",
"Weijia",
""
]
] |
TITLE: FGS-SLAM: Fourier-based Gaussian Splatting for Real-time SLAM with
Sparse and Dense Map Fusion
ABSTRACT: 3D gaussian splatting has advanced simultaneous localization and mapping
(SLAM) technology by enabling real-time positioning and the construction of
high-fidelity maps. However, the uncertainty in gaussian position and
initialization parameters introduces challenges, often requiring extensive
iterative convergence and resulting in redundant or insufficient gaussian
representations. To address this, we introduce a novel adaptive densification
method based on Fourier frequency domain analysis to establish gaussian priors
for rapid convergence. Additionally, we propose constructing independent and
unified sparse and dense maps, where a sparse map supports efficient tracking
via Generalized Iterative Closest Point (GICP) and a dense map creates
high-fidelity visual representations. This is the first SLAM system leveraging
frequency domain analysis to achieve high-quality gaussian mapping in
real-time. Experimental results demonstrate an average frame rate of 36 FPS on
Replica and TUM RGB-D datasets, achieving competitive accuracy in both
localization and mapping.
|
no_new_dataset
| 0.946794 |
2503.01114
|
Junsong Zhang
|
Junsong Zhang, Chunyu Lin, Zhijie Shen, Lang Nie, Kang Liao, Yao Zhao
|
Semi-Supervised 360 Layout Estimation with Panoramic Collaborative
Perturbations
|
9 pages,4 figures
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The performance of existing supervised layout estimation methods heavily
relies on the quality of data annotations. However, obtaining large-scale and
high-quality datasets remains a laborious and time-consuming challenge. To
solve this problem, semi-supervised approaches are introduced to relieve the
demand for expensive data annotations by encouraging the consistent results of
unlabeled data with different perturbations. However, existing solutions merely
employ vanilla perturbations, ignoring the characteristics of panoramic layout
estimation. In contrast, we propose a novel semi-supervised method named
SemiLayout360, which incorporates the priors of the panoramic layout and
distortion through collaborative perturbations. Specifically, we leverage the
panoramic layout prior to enhance the model's focus on potential layout
boundaries. Meanwhile, we introduce the panoramic distortion prior to
strengthen distortion awareness. Furthermore, to prevent intense perturbations
from hindering model convergence and ensure the effectiveness of prior-based
perturbations, we divide and reorganize them as panoramic collaborative
perturbations. Our experimental results on three mainstream benchmarks
demonstrate that the proposed method offers significant advantages over
existing state-of-the-art (SoTA) solutions.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 02:49:20 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhang",
"Junsong",
""
],
[
"Lin",
"Chunyu",
""
],
[
"Shen",
"Zhijie",
""
],
[
"Nie",
"Lang",
""
],
[
"Liao",
"Kang",
""
],
[
"Zhao",
"Yao",
""
]
] |
TITLE: Semi-Supervised 360 Layout Estimation with Panoramic Collaborative
Perturbations
ABSTRACT: The performance of existing supervised layout estimation methods heavily
relies on the quality of data annotations. However, obtaining large-scale and
high-quality datasets remains a laborious and time-consuming challenge. To
solve this problem, semi-supervised approaches are introduced to relieve the
demand for expensive data annotations by encouraging the consistent results of
unlabeled data with different perturbations. However, existing solutions merely
employ vanilla perturbations, ignoring the characteristics of panoramic layout
estimation. In contrast, we propose a novel semi-supervised method named
SemiLayout360, which incorporates the priors of the panoramic layout and
distortion through collaborative perturbations. Specifically, we leverage the
panoramic layout prior to enhance the model's focus on potential layout
boundaries. Meanwhile, we introduce the panoramic distortion prior to
strengthen distortion awareness. Furthermore, to prevent intense perturbations
from hindering model convergence and ensure the effectiveness of prior-based
perturbations, we divide and reorganize them as panoramic collaborative
perturbations. Our experimental results on three mainstream benchmarks
demonstrate that the proposed method offers significant advantages over
existing state-of-the-art (SoTA) solutions.
|
no_new_dataset
| 0.947137 |
2503.01124
|
Shreyas S
|
Shreyas S, Akshath M
|
ViKANformer: Embedding Kolmogorov Arnold Networks in Vision Transformers
for Pattern-Based Learning
|
This paper represents ongoing research and may be subject to
revisions, refinements, and additional experiments in future updates
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Vision Transformers (ViTs) have significantly advanced image classification
by applying self-attention on patch embeddings. However, the standard MLP
blocks in each Transformer layer may not capture complex nonlinear dependencies
optimally. In this paper, we propose ViKANformer, a Vision Transformer where we
replace the MLP sub-layers with Kolmogorov-Arnold Network (KAN) expansions,
including Vanilla KAN, Efficient-KAN, Fast-KAN, SineKAN, and FourierKAN, while
also examining a Flash Attention variant. By leveraging the Kolmogorov-Arnold
theorem, which guarantees that multivariate continuous functions can be
expressed via sums of univariate continuous functions, we aim to boost
representational power. Experimental results on MNIST demonstrate that SineKAN,
Fast-KAN, and a well-tuned Vanilla KAN can achieve over 97% accuracy, albeit
with increased training overhead. This trade-off highlights that KAN expansions
may be beneficial if computational cost is acceptable. We detail the
expansions, present training/test accuracy and F1/ROC metrics, and provide
pseudocode and hyperparameters for reproducibility. Finally, we compare
ViKANformer to a simple MLP and a small CNN baseline on MNIST, illustrating the
efficiency of Transformer-based methods even on a small-scale dataset.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 03:10:26 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"S",
"Shreyas",
""
],
[
"M",
"Akshath",
""
]
] |
TITLE: ViKANformer: Embedding Kolmogorov Arnold Networks in Vision Transformers
for Pattern-Based Learning
ABSTRACT: Vision Transformers (ViTs) have significantly advanced image classification
by applying self-attention on patch embeddings. However, the standard MLP
blocks in each Transformer layer may not capture complex nonlinear dependencies
optimally. In this paper, we propose ViKANformer, a Vision Transformer where we
replace the MLP sub-layers with Kolmogorov-Arnold Network (KAN) expansions,
including Vanilla KAN, Efficient-KAN, Fast-KAN, SineKAN, and FourierKAN, while
also examining a Flash Attention variant. By leveraging the Kolmogorov-Arnold
theorem, which guarantees that multivariate continuous functions can be
expressed via sums of univariate continuous functions, we aim to boost
representational power. Experimental results on MNIST demonstrate that SineKAN,
Fast-KAN, and a well-tuned Vanilla KAN can achieve over 97% accuracy, albeit
with increased training overhead. This trade-off highlights that KAN expansions
may be beneficial if computational cost is acceptable. We detail the
expansions, present training/test accuracy and F1/ROC metrics, and provide
pseudocode and hyperparameters for reproducibility. Finally, we compare
ViKANformer to a simple MLP and a small CNN baseline on MNIST, illustrating the
efficiency of Transformer-based methods even on a small-scale dataset.
|
no_new_dataset
| 0.953188 |
2503.01127
|
Mingao Tan
|
Mingao Tan, Shanze Wang, Biao Huang, Zhibo Yang, Rongfei Chen, Xiaoyu
Shen, Wei Zhang
|
Beyond Visibility Limits: A DRL-Based Navigation Strategy for Unexpected
Obstacles
| null | null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Distance-based reward mechanisms in deep reinforcement learning (DRL)
navigation systems suffer from critical safety limitations in dynamic
environments, frequently resulting in collisions when visibility is restricted.
We propose DRL-NSUO, a novel navigation strategy for unexpected obstacles that
leverages the rate of change in LiDAR data as a dynamic environmental
perception element. Our approach incorporates a composite reward function with
environmental change rate constraints and dynamically adjusted weights through
curriculum learning, enabling robots to autonomously balance between path
efficiency and safety maximization. We enhance sensitivity to nearby obstacles
by implementing short-range feature preprocessing of LiDAR data. Experimental
results demonstrate that this method significantly improves both robot and
pedestrian safety in complex scenarios compared to traditional DRL-based
methods. When evaluated on the BARN navigation dataset, our method achieved
superior performance with success rates of 94.0% at 0.5 m/s and 91.0% at 1.0
m/s, outperforming conservative obstacle expansion strategies. These results
validate DRL-NSUO's enhanced practicality and safety for human-robot
collaborative environments, including intelligent logistics applications.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 03:14:08 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Tan",
"Mingao",
""
],
[
"Wang",
"Shanze",
""
],
[
"Huang",
"Biao",
""
],
[
"Yang",
"Zhibo",
""
],
[
"Chen",
"Rongfei",
""
],
[
"Shen",
"Xiaoyu",
""
],
[
"Zhang",
"Wei",
""
]
] |
TITLE: Beyond Visibility Limits: A DRL-Based Navigation Strategy for Unexpected
Obstacles
ABSTRACT: Distance-based reward mechanisms in deep reinforcement learning (DRL)
navigation systems suffer from critical safety limitations in dynamic
environments, frequently resulting in collisions when visibility is restricted.
We propose DRL-NSUO, a novel navigation strategy for unexpected obstacles that
leverages the rate of change in LiDAR data as a dynamic environmental
perception element. Our approach incorporates a composite reward function with
environmental change rate constraints and dynamically adjusted weights through
curriculum learning, enabling robots to autonomously balance between path
efficiency and safety maximization. We enhance sensitivity to nearby obstacles
by implementing short-range feature preprocessing of LiDAR data. Experimental
results demonstrate that this method significantly improves both robot and
pedestrian safety in complex scenarios compared to traditional DRL-based
methods. When evaluated on the BARN navigation dataset, our method achieved
superior performance with success rates of 94.0% at 0.5 m/s and 91.0% at 1.0
m/s, outperforming conservative obstacle expansion strategies. These results
validate DRL-NSUO's enhanced practicality and safety for human-robot
collaborative environments, including intelligent logistics applications.
|
no_new_dataset
| 0.949106 |
2503.01131
|
Shivam Ratnakar
|
Shivam Ratnakar, Abhiroop Talasila, Raghav Chamadiya, Nikhil Agarwal,
Vinayak K Doifode
|
Beyond QA Pairs: Assessing Parameter-Efficient Fine-Tuning for Fact
Embedding in LLMs
|
Presented at the Workshop on Preparing Good Data for Generative AI:
Challenges and Approaches (Good-Data) in conjunction with AAAI 2025. The
authors retain the copyright
|
Workshop on Preparing Good Data for Generative AI: Challenges and
Approaches, 2025
| null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
This paper presents an extensive examination of Parameter-Efficient
Fine-Tuning (PEFT) for embedding domain specific facts into Large Language
Models (LLMs), focusing on improving the fine-tuning process by categorizing
question-answer (QA) pairs into Factual and Conceptual classes using a
BERT-based classifier. Two distinct Llama-2 models are fine-tuned based on
these classifications and evaluated using larger models like GPT-3.5 Turbo and
Gemini. Our results indicate that models trained on conceptual datasets
outperform those trained on factual datasets. Additionally, we compare the
efficiency of two synthetic fine-tuning dataset generation techniques, D-RAG
and D-Naive, with D-Naive demonstrating superior performance. Although PEFT has
shown effectiveness, our research indicates that it may not be the most optimal
method for embedding facts into LLMs. However, it has demonstrated exceptional
performance in instruction-based tasks. Our findings are reinforced by a
1000-sample dataset in the data center domain, where the fine-tuned Llama-2 7B
model significantly outperforms the baseline model in generating product
recommendations. Our study highlights the importance of QA pair categorization
and synthetic dataset generation techniques in enhancing the performance of
LLMs in specific domains.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 03:26:30 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Ratnakar",
"Shivam",
""
],
[
"Talasila",
"Abhiroop",
""
],
[
"Chamadiya",
"Raghav",
""
],
[
"Agarwal",
"Nikhil",
""
],
[
"Doifode",
"Vinayak K",
""
]
] |
TITLE: Beyond QA Pairs: Assessing Parameter-Efficient Fine-Tuning for Fact
Embedding in LLMs
ABSTRACT: This paper presents an extensive examination of Parameter-Efficient
Fine-Tuning (PEFT) for embedding domain specific facts into Large Language
Models (LLMs), focusing on improving the fine-tuning process by categorizing
question-answer (QA) pairs into Factual and Conceptual classes using a
BERT-based classifier. Two distinct Llama-2 models are fine-tuned based on
these classifications and evaluated using larger models like GPT-3.5 Turbo and
Gemini. Our results indicate that models trained on conceptual datasets
outperform those trained on factual datasets. Additionally, we compare the
efficiency of two synthetic fine-tuning dataset generation techniques, D-RAG
and D-Naive, with D-Naive demonstrating superior performance. Although PEFT has
shown effectiveness, our research indicates that it may not be the most optimal
method for embedding facts into LLMs. However, it has demonstrated exceptional
performance in instruction-based tasks. Our findings are reinforced by a
1000-sample dataset in the data center domain, where the fine-tuned Llama-2 7B
model significantly outperforms the baseline model in generating product
recommendations. Our study highlights the importance of QA pair categorization
and synthetic dataset generation techniques in enhancing the performance of
LLMs in specific domains.
|
no_new_dataset
| 0.949295 |
2503.01144
|
Zhenqi Dai
|
Zhenqi Dai, Ting Liu, Xingxing Zhang, Yunchao Wei, Yanning Zhang
|
One-shot In-context Part Segmentation
|
10 pages
| null |
10.1145/3664647.3680989
| null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we present the One-shot In-context Part Segmentation (OIParts)
framework, designed to tackle the challenges of part segmentation by leveraging
visual foundation models (VFMs). Existing training-based one-shot part
segmentation methods that utilize VFMs encounter difficulties when faced with
scenarios where the one-shot image and test image exhibit significant variance
in appearance and perspective, or when the object in the test image is
partially visible. We argue that training on the one-shot example often leads
to overfitting, thereby compromising the model's generalization capability. Our
framework offers a novel approach to part segmentation that is training-free,
flexible, and data-efficient, requiring only a single in-context example for
precise segmentation with superior generalization ability. By thoroughly
exploring the complementary strengths of VFMs, specifically DINOv2 and Stable
Diffusion, we introduce an adaptive channel selection approach by minimizing
the intra-class distance for better exploiting these two features, thereby
enhancing the discriminatory power of the extracted features for the
fine-grained parts. We have achieved remarkable segmentation performance across
diverse object categories. The OIParts framework not only eliminates the need
for extensive labeled data but also demonstrates superior generalization
ability. Through comprehensive experimentation on three benchmark datasets, we
have demonstrated the superiority of our proposed method over existing part
segmentation approaches in one-shot settings.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 03:50:54 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Dai",
"Zhenqi",
""
],
[
"Liu",
"Ting",
""
],
[
"Zhang",
"Xingxing",
""
],
[
"Wei",
"Yunchao",
""
],
[
"Zhang",
"Yanning",
""
]
] |
TITLE: One-shot In-context Part Segmentation
ABSTRACT: In this paper, we present the One-shot In-context Part Segmentation (OIParts)
framework, designed to tackle the challenges of part segmentation by leveraging
visual foundation models (VFMs). Existing training-based one-shot part
segmentation methods that utilize VFMs encounter difficulties when faced with
scenarios where the one-shot image and test image exhibit significant variance
in appearance and perspective, or when the object in the test image is
partially visible. We argue that training on the one-shot example often leads
to overfitting, thereby compromising the model's generalization capability. Our
framework offers a novel approach to part segmentation that is training-free,
flexible, and data-efficient, requiring only a single in-context example for
precise segmentation with superior generalization ability. By thoroughly
exploring the complementary strengths of VFMs, specifically DINOv2 and Stable
Diffusion, we introduce an adaptive channel selection approach by minimizing
the intra-class distance for better exploiting these two features, thereby
enhancing the discriminatory power of the extracted features for the
fine-grained parts. We have achieved remarkable segmentation performance across
diverse object categories. The OIParts framework not only eliminates the need
for extensive labeled data but also demonstrates superior generalization
ability. Through comprehensive experimentation on three benchmark datasets, we
have demonstrated the superiority of our proposed method over existing part
segmentation approaches in one-shot settings.
|
no_new_dataset
| 0.950457 |
2503.01152
|
Difan Zou
|
Shilin Tong, Difei Wu, Xiaona Liu, Le Zheng, Yuchuan Du, Difan Zou
|
STGAN: Spatial-temporal Graph Autoregression Network for Pavement
Distress Deterioration Prediction
|
16 pages, 16 figures, 4 tables, accepted by IEEE Transactions on
Intelligent Transportation Systems (TITS)
| null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Pavement distress significantly compromises road integrity and poses risks to
drivers. Accurate prediction of pavement distress deterioration is essential
for effective road management, cost reduction in maintenance, and improvement
of traffic safety. However, real-world data on pavement distress is usually
collected irregularly, resulting in uneven, asynchronous, and sparse
spatial-temporal datasets. This hinders the application of existing
spatial-temporal models, such as DCRNN, since they are only applicable to
regularly and synchronously collected data. To overcome these challenges, we
propose the Spatial-Temporal Graph Autoregression Network (STGAN), a novel
graph neural network model designed for accurately predicting irregular
pavement distress deterioration using complex spatial-temporal data.
Specifically, STGAN integrates the temporal domain into the spatial domain,
creating a larger graph where nodes are represented by spatial-temporal tuples
and edges are formed based on a similarity-based connection mechanism.
Furthermore, based on the constructed spatiotemporal graph, we formulate
pavement distress deterioration prediction as a graph autoregression task,
i.e., the graph size increases incrementally and the prediction is performed
sequentially. This is accomplished by a novel spatial-temporal attention
mechanism deployed by STGAN. Utilizing the ConTrack dataset, which contains
pavement distress records collected from different locations in Shanghai, we
demonstrate the superior performance of STGAN in capturing spatial-temporal
correlations and addressing the aforementioned challenges. Experimental results
further show that STGAN outperforms baseline models, and ablation studies
confirm the effectiveness of its novel modules. Our findings contribute to
promoting proactive road maintenance decision-making and ultimately enhancing
road safety and resilience.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 03:59:34 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Tong",
"Shilin",
""
],
[
"Wu",
"Difei",
""
],
[
"Liu",
"Xiaona",
""
],
[
"Zheng",
"Le",
""
],
[
"Du",
"Yuchuan",
""
],
[
"Zou",
"Difan",
""
]
] |
TITLE: STGAN: Spatial-temporal Graph Autoregression Network for Pavement
Distress Deterioration Prediction
ABSTRACT: Pavement distress significantly compromises road integrity and poses risks to
drivers. Accurate prediction of pavement distress deterioration is essential
for effective road management, cost reduction in maintenance, and improvement
of traffic safety. However, real-world data on pavement distress is usually
collected irregularly, resulting in uneven, asynchronous, and sparse
spatial-temporal datasets. This hinders the application of existing
spatial-temporal models, such as DCRNN, since they are only applicable to
regularly and synchronously collected data. To overcome these challenges, we
propose the Spatial-Temporal Graph Autoregression Network (STGAN), a novel
graph neural network model designed for accurately predicting irregular
pavement distress deterioration using complex spatial-temporal data.
Specifically, STGAN integrates the temporal domain into the spatial domain,
creating a larger graph where nodes are represented by spatial-temporal tuples
and edges are formed based on a similarity-based connection mechanism.
Furthermore, based on the constructed spatiotemporal graph, we formulate
pavement distress deterioration prediction as a graph autoregression task,
i.e., the graph size increases incrementally and the prediction is performed
sequentially. This is accomplished by a novel spatial-temporal attention
mechanism deployed by STGAN. Utilizing the ConTrack dataset, which contains
pavement distress records collected from different locations in Shanghai, we
demonstrate the superior performance of STGAN in capturing spatial-temporal
correlations and addressing the aforementioned challenges. Experimental results
further show that STGAN outperforms baseline models, and ablation studies
confirm the effectiveness of its novel modules. Our findings contribute to
promoting proactive road maintenance decision-making and ultimately enhancing
road safety and resilience.
|
no_new_dataset
| 0.941708 |
2503.01158
|
Lincheng Li
|
Suzhen Wang, Weijie Chen, Wei Zhang, Minda Zhao, Lincheng Li,
Rongsheng Zhang, Zhipeng Hu, Xin Yu
|
EasyCraft: A Robust and Efficient Framework for Automatic Avatar
Crafting
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Character customization, or 'face crafting,' is a vital feature in
role-playing games (RPGs), enhancing player engagement by enabling the creation
of personalized avatars. Existing automated methods often struggle with
generalizability across diverse game engines due to their reliance on the
intermediate constraints of specific image domain and typically support only
one type of input, either text or image. To overcome these challenges, we
introduce EasyCraft, an innovative end-to-end feedforward framework that
automates character crafting by uniquely supporting both text and image inputs.
Our approach employs a translator capable of converting facial images of any
style into crafting parameters. We first establish a unified feature
distribution in the translator's image encoder through self-supervised learning
on a large-scale dataset, enabling photos of any style to be embedded into a
unified feature representation. Subsequently, we map this unified feature
distribution to crafting parameters specific to a game engine, a process that
can be easily adapted to most game engines and thus enhances EasyCraft's
generalizability. By integrating text-to-image techniques with our translator,
EasyCraft also facilitates precise, text-based character crafting. EasyCraft's
ability to integrate diverse inputs significantly enhances the versatility and
accuracy of avatar creation. Extensive experiments on two RPG games demonstrate
the effectiveness of our method, achieving state-of-the-art results and
facilitating adaptability across various avatar engines.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 04:11:47 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Wang",
"Suzhen",
""
],
[
"Chen",
"Weijie",
""
],
[
"Zhang",
"Wei",
""
],
[
"Zhao",
"Minda",
""
],
[
"Li",
"Lincheng",
""
],
[
"Zhang",
"Rongsheng",
""
],
[
"Hu",
"Zhipeng",
""
],
[
"Yu",
"Xin",
""
]
] |
TITLE: EasyCraft: A Robust and Efficient Framework for Automatic Avatar
Crafting
ABSTRACT: Character customization, or 'face crafting,' is a vital feature in
role-playing games (RPGs), enhancing player engagement by enabling the creation
of personalized avatars. Existing automated methods often struggle with
generalizability across diverse game engines due to their reliance on the
intermediate constraints of specific image domain and typically support only
one type of input, either text or image. To overcome these challenges, we
introduce EasyCraft, an innovative end-to-end feedforward framework that
automates character crafting by uniquely supporting both text and image inputs.
Our approach employs a translator capable of converting facial images of any
style into crafting parameters. We first establish a unified feature
distribution in the translator's image encoder through self-supervised learning
on a large-scale dataset, enabling photos of any style to be embedded into a
unified feature representation. Subsequently, we map this unified feature
distribution to crafting parameters specific to a game engine, a process that
can be easily adapted to most game engines and thus enhances EasyCraft's
generalizability. By integrating text-to-image techniques with our translator,
EasyCraft also facilitates precise, text-based character crafting. EasyCraft's
ability to integrate diverse inputs significantly enhances the versatility and
accuracy of avatar creation. Extensive experiments on two RPG games demonstrate
the effectiveness of our method, achieving state-of-the-art results and
facilitating adaptability across various avatar engines.
|
no_new_dataset
| 0.945147 |
2503.01164
|
Yitao Zhu
|
Yitao Zhu, Yuan Yin, Jiaming Li, Mengjie Xu, Zihao Zhao, Honglin
Xiong, Sheng Wang, Qian Wang
|
Med-LEGO: Editing and Adapting toward Generalist Medical Image Diagnosis
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The adoption of visual foundation models has become a common practice in
computer-aided diagnosis (CAD). While these foundation models provide a viable
solution for creating generalist medical AI, privacy concerns make it difficult
to pre-train or continuously update such models across multiple domains and
datasets, leading many studies to focus on specialist models. To address this
challenge, we propose Med-LEGO, a training-free framework that enables the
seamless integration or updating of a generalist CAD model by combining
multiple specialist models, similar to assembling LEGO bricks. Med-LEGO
enhances LoRA (low-rank adaptation) by incorporating singular value
decomposition (SVD) to efficiently capture the domain expertise of each
specialist model with minimal additional parameters. By combining these adapted
weights through simple operations, Med-LEGO allows for the easy integration or
modification of specific diagnostic capabilities without the need for original
data or retraining. Finally, the combined model can be further adapted to new
diagnostic tasks, making it a versatile generalist model. Our extensive
experiments demonstrate that Med-LEGO outperforms existing methods in both
cross-domain and in-domain medical tasks while using only 0.18% of full model
parameters. These merged models show better convergence and generalization to
new tasks, providing an effective path toward generalist medical AI.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 04:27:11 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhu",
"Yitao",
""
],
[
"Yin",
"Yuan",
""
],
[
"Li",
"Jiaming",
""
],
[
"Xu",
"Mengjie",
""
],
[
"Zhao",
"Zihao",
""
],
[
"Xiong",
"Honglin",
""
],
[
"Wang",
"Sheng",
""
],
[
"Wang",
"Qian",
""
]
] |
TITLE: Med-LEGO: Editing and Adapting toward Generalist Medical Image Diagnosis
ABSTRACT: The adoption of visual foundation models has become a common practice in
computer-aided diagnosis (CAD). While these foundation models provide a viable
solution for creating generalist medical AI, privacy concerns make it difficult
to pre-train or continuously update such models across multiple domains and
datasets, leading many studies to focus on specialist models. To address this
challenge, we propose Med-LEGO, a training-free framework that enables the
seamless integration or updating of a generalist CAD model by combining
multiple specialist models, similar to assembling LEGO bricks. Med-LEGO
enhances LoRA (low-rank adaptation) by incorporating singular value
decomposition (SVD) to efficiently capture the domain expertise of each
specialist model with minimal additional parameters. By combining these adapted
weights through simple operations, Med-LEGO allows for the easy integration or
modification of specific diagnostic capabilities without the need for original
data or retraining. Finally, the combined model can be further adapted to new
diagnostic tasks, making it a versatile generalist model. Our extensive
experiments demonstrate that Med-LEGO outperforms existing methods in both
cross-domain and in-domain medical tasks while using only 0.18% of full model
parameters. These merged models show better convergence and generalization to
new tasks, providing an effective path toward generalist medical AI.
|
no_new_dataset
| 0.948106 |
2503.01169
|
Seyed Mohamad Ali Tousi
|
Seyed Mohamad Ali Tousi, Ramy Farag, Jacket Demby's, Gbenga Omotara,
John A. Lory, G. N. DeSouza
|
A Zero-Shot Learning Approach for Ephemeral Gully Detection from Remote
Sensing using Vision Language Models
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Ephemeral gullies are a primary cause of soil erosion and their reliable,
accurate, and early detection will facilitate significant improvements in the
sustainability of global agricultural systems. In our view, prior research has
not successfully addressed automated detection of ephemeral gullies from
remotely sensed images, so for the first time, we present and evaluate three
successful pipelines for ephemeral gully detection. Our pipelines utilize
remotely sensed images, acquired from specific agricultural areas over a period
of time. The pipelines were tested with various choices of Visual Language
Models (VLMs), and they classified the images based on the presence of
ephemeral gullies with accuracy higher than 70% and a F1-score close to 80% for
positive gully detection. Additionally, we developed the first public dataset
for ephemeral gully detection, labeled by a team of soil- and plant-science
experts. To evaluate the proposed pipelines, we employed a variety of zero-shot
classification methods based on State-of-the-Art (SOTA) open-source
Vision-Language Models (VLMs). In addition to that, we compare the same
pipelines with a transfer learning approach. Extensive experiments were
conducted to validate the detection pipelines and to analyze the impact of
hyperparameter changes in their performance. The experimental results
demonstrate that the proposed zero-shot classification pipelines are highly
effective in detecting ephemeral gullies in a scenario where classification
datasets are scarce.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 04:36:25 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Tousi",
"Seyed Mohamad Ali",
""
],
[
"Farag",
"Ramy",
""
],
[
"Demby's",
"Jacket",
""
],
[
"Omotara",
"Gbenga",
""
],
[
"Lory",
"John A.",
""
],
[
"DeSouza",
"G. N.",
""
]
] |
TITLE: A Zero-Shot Learning Approach for Ephemeral Gully Detection from Remote
Sensing using Vision Language Models
ABSTRACT: Ephemeral gullies are a primary cause of soil erosion and their reliable,
accurate, and early detection will facilitate significant improvements in the
sustainability of global agricultural systems. In our view, prior research has
not successfully addressed automated detection of ephemeral gullies from
remotely sensed images, so for the first time, we present and evaluate three
successful pipelines for ephemeral gully detection. Our pipelines utilize
remotely sensed images, acquired from specific agricultural areas over a period
of time. The pipelines were tested with various choices of Visual Language
Models (VLMs), and they classified the images based on the presence of
ephemeral gullies with accuracy higher than 70% and a F1-score close to 80% for
positive gully detection. Additionally, we developed the first public dataset
for ephemeral gully detection, labeled by a team of soil- and plant-science
experts. To evaluate the proposed pipelines, we employed a variety of zero-shot
classification methods based on State-of-the-Art (SOTA) open-source
Vision-Language Models (VLMs). In addition to that, we compare the same
pipelines with a transfer learning approach. Extensive experiments were
conducted to validate the detection pipelines and to analyze the impact of
hyperparameter changes in their performance. The experimental results
demonstrate that the proposed zero-shot classification pipelines are highly
effective in detecting ephemeral gullies in a scenario where classification
datasets are scarce.
|
new_dataset
| 0.959116 |
2503.01176
|
Kart-Leong Lim
|
Kart-Leong Lim, Rahul Dutta
|
Prognostics and Health Management of Wafer Chemical-Mechanical Polishing
System using Autoencoder
| null | null | null | null |
cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The Prognostics and Health Management Data Challenge (PHM) 2016 tracks the
health state of components of a semiconductor wafer polishing process. The
ultimate goal is to develop an ability to predict the measurement on the wafer
surface wear through monitoring the components health state. This translates to
cost saving in large scale production. The PHM dataset contains many time
series measurements not utilized by traditional physics based approach. On the
other hand task, applying a data driven approach such as deep learning to the
PHM dataset is non-trivial. The main issue with supervised deep learning is
that class label is not available to the PHM dataset. Second, the feature space
trained by an unsupervised deep learner is not specifically targeted at the
predictive ability or regression. In this work, we propose using the
autoencoder based clustering whereby the feature space trained is found to be
more suitable for performing regression. This is due to having a more compact
distribution of samples respective to their nearest cluster means. We justify
our claims by comparing the performance of our proposed method on the PHM
dataset with several baselines such as the autoencoder as well as
state-of-the-art approaches.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 04:48:34 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Lim",
"Kart-Leong",
""
],
[
"Dutta",
"Rahul",
""
]
] |
TITLE: Prognostics and Health Management of Wafer Chemical-Mechanical Polishing
System using Autoencoder
ABSTRACT: The Prognostics and Health Management Data Challenge (PHM) 2016 tracks the
health state of components of a semiconductor wafer polishing process. The
ultimate goal is to develop an ability to predict the measurement on the wafer
surface wear through monitoring the components health state. This translates to
cost saving in large scale production. The PHM dataset contains many time
series measurements not utilized by traditional physics based approach. On the
other hand task, applying a data driven approach such as deep learning to the
PHM dataset is non-trivial. The main issue with supervised deep learning is
that class label is not available to the PHM dataset. Second, the feature space
trained by an unsupervised deep learner is not specifically targeted at the
predictive ability or regression. In this work, we propose using the
autoencoder based clustering whereby the feature space trained is found to be
more suitable for performing regression. This is due to having a more compact
distribution of samples respective to their nearest cluster means. We justify
our claims by comparing the performance of our proposed method on the PHM
dataset with several baselines such as the autoencoder as well as
state-of-the-art approaches.
|
no_new_dataset
| 0.943191 |
2503.01184
|
EungGu Yun
|
EungGu Yun, Heonjin Ha, Yeongwoo Nam, Bryan Dongik Lee
|
Language-Assisted Feature Transformation for Anomaly Detection
|
ICLR 2025
| null | null | null |
cs.LG cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper introduces LAFT, a novel feature transformation method designed to
incorporate user knowledge and preferences into anomaly detection using natural
language. Accurately modeling the boundary of normality is crucial for
distinguishing abnormal data, but this is often challenging due to limited data
or the presence of nuisance attributes. While unsupervised methods that rely
solely on data without user guidance are common, they may fail to detect
anomalies of specific interest. To address this limitation, we propose
Language-Assisted Feature Transformation (LAFT), which leverages the shared
image-text embedding space of vision-language models to transform visual
features according to user-defined requirements. Combined with anomaly
detection methods, LAFT effectively aligns visual features with user
preferences, allowing anomalies of interest to be detected. Extensive
experiments on both toy and real-world datasets validate the effectiveness of
our method.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 05:15:49 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Yun",
"EungGu",
""
],
[
"Ha",
"Heonjin",
""
],
[
"Nam",
"Yeongwoo",
""
],
[
"Lee",
"Bryan Dongik",
""
]
] |
TITLE: Language-Assisted Feature Transformation for Anomaly Detection
ABSTRACT: This paper introduces LAFT, a novel feature transformation method designed to
incorporate user knowledge and preferences into anomaly detection using natural
language. Accurately modeling the boundary of normality is crucial for
distinguishing abnormal data, but this is often challenging due to limited data
or the presence of nuisance attributes. While unsupervised methods that rely
solely on data without user guidance are common, they may fail to detect
anomalies of specific interest. To address this limitation, we propose
Language-Assisted Feature Transformation (LAFT), which leverages the shared
image-text embedding space of vision-language models to transform visual
features according to user-defined requirements. Combined with anomaly
detection methods, LAFT effectively aligns visual features with user
preferences, allowing anomalies of interest to be detected. Extensive
experiments on both toy and real-world datasets validate the effectiveness of
our method.
|
no_new_dataset
| 0.952926 |
2503.01199
|
Kaimin Liao
|
Kaimin Liao
|
LiteGS: A High-Performance Modular Framework for Gaussian Splatting
Training
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Gaussian splatting has emerged as a powerful technique for reconstruction of
3D scenes in computer graphics and vision. However, conventional
implementations often suffer from inefficiencies, limited flexibility, and high
computational overhead, which constrain their adaptability to diverse
applications. In this paper, we present LiteGS,a high-performance and modular
framework that enhances both the efficiency and usability of Gaussian
splatting. LiteGS achieves a 3.4x speedup over the original 3DGS implementation
while reducing GPU memory usage by approximately 30%. Its modular design
decomposes the splatting process into multiple highly optimized operators, and
it provides dual API support via a script-based interface and a CUDA-based
interface. The script-based interface, in combination with autograd, enables
rapid prototyping and straightforward customization of new ideas, while the
CUDA-based interface delivers optimal training speeds for performance-critical
applications. LiteGS retains the core algorithm of 3DGS, ensuring
compatibility. Comprehensive experiments on the Mip-NeRF 360 dataset
demonstrate that LiteGS accelerates training without compromising accuracy,
making it an ideal solution for both rapid prototyping and production
environments.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 05:52:02 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Liao",
"Kaimin",
""
]
] |
TITLE: LiteGS: A High-Performance Modular Framework for Gaussian Splatting
Training
ABSTRACT: Gaussian splatting has emerged as a powerful technique for reconstruction of
3D scenes in computer graphics and vision. However, conventional
implementations often suffer from inefficiencies, limited flexibility, and high
computational overhead, which constrain their adaptability to diverse
applications. In this paper, we present LiteGS,a high-performance and modular
framework that enhances both the efficiency and usability of Gaussian
splatting. LiteGS achieves a 3.4x speedup over the original 3DGS implementation
while reducing GPU memory usage by approximately 30%. Its modular design
decomposes the splatting process into multiple highly optimized operators, and
it provides dual API support via a script-based interface and a CUDA-based
interface. The script-based interface, in combination with autograd, enables
rapid prototyping and straightforward customization of new ideas, while the
CUDA-based interface delivers optimal training speeds for performance-critical
applications. LiteGS retains the core algorithm of 3DGS, ensuring
compatibility. Comprehensive experiments on the Mip-NeRF 360 dataset
demonstrate that LiteGS accelerates training without compromising accuracy,
making it an ideal solution for both rapid prototyping and production
environments.
|
no_new_dataset
| 0.940408 |
2503.01203
|
Yifan Feng
|
Yifan Feng, Shiquan Liu, Xiangmin Han, Shaoyi Du, Zongze Wu, Han Hu,
Yue Gao
|
Hypergraph Foundation Model
| null | null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Hypergraph neural networks (HGNNs) effectively model complex high-order
relationships in domains like protein interactions and social networks by
connecting multiple vertices through hyperedges, enhancing modeling
capabilities, and reducing information loss. Developing foundation models for
hypergraphs is challenging due to their distinct data, which includes both
vertex features and intricate structural information. We present Hyper-FM, a
Hypergraph Foundation Model for multi-domain knowledge extraction, featuring
Hierarchical High-Order Neighbor Guided Vertex Knowledge Embedding for vertex
feature representation and Hierarchical Multi-Hypergraph Guided Structural
Knowledge Extraction for structural information. Additionally, we curate 10
text-attributed hypergraph datasets to advance research between HGNNs and LLMs.
Experiments on these datasets show that Hyper-FM outperforms baseline methods
by approximately 13.3\%, validating our approach. Furthermore, we propose the
first scaling law for hypergraph foundation models, demonstrating that
increasing domain diversity significantly enhances performance, unlike merely
augmenting vertex and hyperedge counts. This underscores the critical role of
domain diversity in scaling hypergraph models.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 05:56:08 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Feng",
"Yifan",
""
],
[
"Liu",
"Shiquan",
""
],
[
"Han",
"Xiangmin",
""
],
[
"Du",
"Shaoyi",
""
],
[
"Wu",
"Zongze",
""
],
[
"Hu",
"Han",
""
],
[
"Gao",
"Yue",
""
]
] |
TITLE: Hypergraph Foundation Model
ABSTRACT: Hypergraph neural networks (HGNNs) effectively model complex high-order
relationships in domains like protein interactions and social networks by
connecting multiple vertices through hyperedges, enhancing modeling
capabilities, and reducing information loss. Developing foundation models for
hypergraphs is challenging due to their distinct data, which includes both
vertex features and intricate structural information. We present Hyper-FM, a
Hypergraph Foundation Model for multi-domain knowledge extraction, featuring
Hierarchical High-Order Neighbor Guided Vertex Knowledge Embedding for vertex
feature representation and Hierarchical Multi-Hypergraph Guided Structural
Knowledge Extraction for structural information. Additionally, we curate 10
text-attributed hypergraph datasets to advance research between HGNNs and LLMs.
Experiments on these datasets show that Hyper-FM outperforms baseline methods
by approximately 13.3\%, validating our approach. Furthermore, we propose the
first scaling law for hypergraph foundation models, demonstrating that
increasing domain diversity significantly enhances performance, unlike merely
augmenting vertex and hyperedge counts. This underscores the critical role of
domain diversity in scaling hypergraph models.
|
no_new_dataset
| 0.884888 |
2503.01212
|
Deyu Bo
|
Deyu Bo, Songhua Liu, Xinchao Wang
|
Understanding Dataset Distillation via Spectral Filtering
| null | null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Dataset distillation (DD) has emerged as a promising approach to compress
datasets and speed up model training. However, the underlying connections among
various DD methods remain largely unexplored. In this paper, we introduce
UniDD, a spectral filtering framework that unifies diverse DD objectives. UniDD
interprets each DD objective as a specific filter function that affects the
eigenvalues of the feature-feature correlation (FFC) matrix and modulates the
frequency components of the feature-label correlation (FLC) matrix. In this
way, UniDD reveals that the essence of DD fundamentally lies in matching
frequency-specific features. Moreover, according to the filter behaviors, we
classify existing methods into low-frequency matching and high-frequency
matching, encoding global texture and local details, respectively. However,
existing methods rely on fixed filter functions throughout distillation, which
cannot capture the low- and high-frequency information simultaneously. To
address this limitation, we further propose Curriculum Frequency Matching
(CFM), which gradually adjusts the filter parameter to cover both low- and
high-frequency information of the FFC and FLC matrices. Extensive experiments
on small-scale datasets, such as CIFAR-10/100, and large-scale datasets,
including ImageNet-1K, demonstrate the superior performance of CFM over
existing baselines and validate the practicality of UniDD.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:22:34 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Bo",
"Deyu",
""
],
[
"Liu",
"Songhua",
""
],
[
"Wang",
"Xinchao",
""
]
] |
TITLE: Understanding Dataset Distillation via Spectral Filtering
ABSTRACT: Dataset distillation (DD) has emerged as a promising approach to compress
datasets and speed up model training. However, the underlying connections among
various DD methods remain largely unexplored. In this paper, we introduce
UniDD, a spectral filtering framework that unifies diverse DD objectives. UniDD
interprets each DD objective as a specific filter function that affects the
eigenvalues of the feature-feature correlation (FFC) matrix and modulates the
frequency components of the feature-label correlation (FLC) matrix. In this
way, UniDD reveals that the essence of DD fundamentally lies in matching
frequency-specific features. Moreover, according to the filter behaviors, we
classify existing methods into low-frequency matching and high-frequency
matching, encoding global texture and local details, respectively. However,
existing methods rely on fixed filter functions throughout distillation, which
cannot capture the low- and high-frequency information simultaneously. To
address this limitation, we further propose Curriculum Frequency Matching
(CFM), which gradually adjusts the filter parameter to cover both low- and
high-frequency information of the FFC and FLC matrices. Extensive experiments
on small-scale datasets, such as CIFAR-10/100, and large-scale datasets,
including ImageNet-1K, demonstrate the superior performance of CFM over
existing baselines and validate the practicality of UniDD.
|
no_new_dataset
| 0.944791 |
2503.01214
|
Yuhan Bao
|
Yuhan Bao, Shaohua Gao, Wenyong Li and Kaiwei Wang
|
One-Step Event-Driven High-Speed Autofocus
|
Main text: 9 pages, 6 figures. Supplementary Material: 4 pages, 3
figures. Accepted by CVPR2025
| null | null | null |
cs.CV physics.optics
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
High-speed autofocus in extreme scenes remains a significant challenge.
Traditional methods rely on repeated sampling around the focus position,
resulting in ``focus hunting''. Event-driven methods have advanced focusing
speed and improved performance in low-light conditions; however, current
approaches still require at least one lengthy round of ``focus hunting'',
involving the collection of a complete focus stack. We introduce the Event
Laplacian Product (ELP) focus detection function, which combines event data
with grayscale Laplacian information, redefining focus search as a detection
task. This innovation enables the first one-step event-driven autofocus,
cutting focusing time by up to two-thirds and reducing focusing error by 24
times on the DAVIS346 dataset and 22 times on the EVK4 dataset. Additionally,
we present an autofocus pipeline tailored for event-only cameras, achieving
accurate results across a range of challenging motion and lighting conditions.
All datasets and code will be made publicly available.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:25:09 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Bao",
"Yuhan",
""
],
[
"Gao",
"Shaohua",
""
],
[
"Li",
"Wenyong",
""
],
[
"Wang",
"Kaiwei",
""
]
] |
TITLE: One-Step Event-Driven High-Speed Autofocus
ABSTRACT: High-speed autofocus in extreme scenes remains a significant challenge.
Traditional methods rely on repeated sampling around the focus position,
resulting in ``focus hunting''. Event-driven methods have advanced focusing
speed and improved performance in low-light conditions; however, current
approaches still require at least one lengthy round of ``focus hunting'',
involving the collection of a complete focus stack. We introduce the Event
Laplacian Product (ELP) focus detection function, which combines event data
with grayscale Laplacian information, redefining focus search as a detection
task. This innovation enables the first one-step event-driven autofocus,
cutting focusing time by up to two-thirds and reducing focusing error by 24
times on the DAVIS346 dataset and 22 times on the EVK4 dataset. Additionally,
we present an autofocus pipeline tailored for event-only cameras, achieving
accurate results across a range of challenging motion and lighting conditions.
All datasets and code will be made publicly available.
|
no_new_dataset
| 0.933975 |
2503.01217
|
Sijin Sun
|
Sijin Sun, Ming Deng, Xinrui Yu, Liangbin Zhao
|
HREB-CRF: Hierarchical Reduced-bias EMA for Chinese Named Entity
Recognition
|
18 pages, 10 figures, under Review
| null | null | null |
cs.CL cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Incorrect boundary division, complex semantic representation, and differences
in pronunciation and meaning often lead to errors in Chinese Named Entity
Recognition(CNER). To address these issues, this paper proposes HREB-CRF
framework: Hierarchical Reduced-bias EMA with CRF. The proposed method
amplifies word boundaries and pools long text gradients through exponentially
fixed-bias weighted average of local and global hierarchical attention.
Experimental results on the MSRA, Resume, and Weibo datasets show excellent in
F1, outperforming the baseline model by 1.1\%, 1.6\%, and 9.8\%. The
significant improvement in F1 shows evidences of strong effectiveness and
robustness of approach in CNER tasks.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:31:52 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Sun",
"Sijin",
""
],
[
"Deng",
"Ming",
""
],
[
"Yu",
"Xinrui",
""
],
[
"Zhao",
"Liangbin",
""
]
] |
TITLE: HREB-CRF: Hierarchical Reduced-bias EMA for Chinese Named Entity
Recognition
ABSTRACT: Incorrect boundary division, complex semantic representation, and differences
in pronunciation and meaning often lead to errors in Chinese Named Entity
Recognition(CNER). To address these issues, this paper proposes HREB-CRF
framework: Hierarchical Reduced-bias EMA with CRF. The proposed method
amplifies word boundaries and pools long text gradients through exponentially
fixed-bias weighted average of local and global hierarchical attention.
Experimental results on the MSRA, Resume, and Weibo datasets show excellent in
F1, outperforming the baseline model by 1.1\%, 1.6\%, and 9.8\%. The
significant improvement in F1 shows evidences of strong effectiveness and
robustness of approach in CNER tasks.
|
no_new_dataset
| 0.952397 |
2503.01221
|
Bas Peters
|
Ophir Greif, Bas Peters, Michael S. McMillan, Paulina Wozniakowska,
Eldad Haber
|
Machine Learning for Airborne Electromagnetic Data Inversion: a
Bootstrapped Approach
|
16 pages, 9 figures
| null | null | null |
physics.geo-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Aircraft-based surveying to collect airborne electromagnetic data is a key
method to image large swaths of the Earth's surface in pursuit of better
knowledge of aquifer systems. Despite many years of advancements, 3D inversion
still poses challenges in terms of computational requirements, regularization
selection, hyperparameter tuning and real-time inversion. We present a new
approach for the inversion of airborne electromagnetic data that leverages
machine learning to overcome the computational burden of traditional 3D
inversion methods, which implicitly includes learned regularization and is
applicable in real-time. The method combines 1D inversion results with
geostatistical modeling to create tailored training datasets, enabling the
development of a specialized neural network that predicts 2D conductivity
models from airborne electromagnetic data. This approach requires 3D forward
modeling and 1D inversion up front, but no forward modeling during inference.
The workflow is applied to the Kaweah Subbasin in California, where it
successfully reconstructs conductivity models consistent with real-world data
and geological drill hole information. The results highlight the method's
capability to deliver fast and accurate subsurface imaging, offering a valuable
tool for groundwater exploration and other near-surface applications.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:39:28 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Greif",
"Ophir",
""
],
[
"Peters",
"Bas",
""
],
[
"McMillan",
"Michael S.",
""
],
[
"Wozniakowska",
"Paulina",
""
],
[
"Haber",
"Eldad",
""
]
] |
TITLE: Machine Learning for Airborne Electromagnetic Data Inversion: a
Bootstrapped Approach
ABSTRACT: Aircraft-based surveying to collect airborne electromagnetic data is a key
method to image large swaths of the Earth's surface in pursuit of better
knowledge of aquifer systems. Despite many years of advancements, 3D inversion
still poses challenges in terms of computational requirements, regularization
selection, hyperparameter tuning and real-time inversion. We present a new
approach for the inversion of airborne electromagnetic data that leverages
machine learning to overcome the computational burden of traditional 3D
inversion methods, which implicitly includes learned regularization and is
applicable in real-time. The method combines 1D inversion results with
geostatistical modeling to create tailored training datasets, enabling the
development of a specialized neural network that predicts 2D conductivity
models from airborne electromagnetic data. This approach requires 3D forward
modeling and 1D inversion up front, but no forward modeling during inference.
The workflow is applied to the Kaweah Subbasin in California, where it
successfully reconstructs conductivity models consistent with real-world data
and geological drill hole information. The results highlight the method's
capability to deliver fast and accurate subsurface imaging, offering a valuable
tool for groundwater exploration and other near-surface applications.
|
no_new_dataset
| 0.947769 |
2503.01226
|
Sahar Sinene Mehdoui
|
Sahar Sinene Mehdoui, Abdelhamid Bouzid, Daniel Sierra-Sosa and Adel
Elmaghraby
|
Dementia Insights: A Context-Based MultiModal Approach
| null | null | null | null |
q-bio.NC cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Dementia, a progressive neurodegenerative disorder, affects memory,
reasoning, and daily functioning, creating challenges for individuals and
healthcare systems. Early detection is crucial for timely interventions that
may slow disease progression. Large pre-trained models (LPMs) for text and
audio, such as Generative Pre-trained Transformer (GPT), Bidirectional Encoder
Representations from Transformers (BERT), and Contrastive Language-Audio
Pretraining (CLAP), have shown promise in identifying cognitive impairments.
However, existing studies generally rely heavily on expert-annotated datasets
and unimodal approaches, limiting robustness and scalability. This study
proposes a context-based multimodal method, integrating both text and audio
data using the best-performing LPMs in each modality. By incorporating
contextual embeddings, our method improves dementia detection performance.
Additionally, motivated by the effectiveness of contextual embeddings, we
further experimented with a context-based In-Context Learning (ICL) as a
complementary technique. Results show that GPT-based embeddings, particularly
when fused with CLAP audio features, achieve an F1-score of $83.33\%$,
surpassing state-of-the-art dementia detection models. Furthermore, raw text
data outperforms expert-annotated datasets, demonstrating that LPMs can extract
meaningful linguistic and acoustic patterns without extensive manual labeling.
These findings highlight the potential for scalable, non-invasive diagnostic
tools that reduce reliance on costly annotations while maintaining high
accuracy. By integrating multimodal learning with contextual embeddings, this
work lays the foundation for future advancements in personalized dementia
detection and cognitive health research.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:46:26 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Mehdoui",
"Sahar Sinene",
""
],
[
"Bouzid",
"Abdelhamid",
""
],
[
"Sierra-Sosa",
"Daniel",
""
],
[
"Elmaghraby",
"Adel",
""
]
] |
TITLE: Dementia Insights: A Context-Based MultiModal Approach
ABSTRACT: Dementia, a progressive neurodegenerative disorder, affects memory,
reasoning, and daily functioning, creating challenges for individuals and
healthcare systems. Early detection is crucial for timely interventions that
may slow disease progression. Large pre-trained models (LPMs) for text and
audio, such as Generative Pre-trained Transformer (GPT), Bidirectional Encoder
Representations from Transformers (BERT), and Contrastive Language-Audio
Pretraining (CLAP), have shown promise in identifying cognitive impairments.
However, existing studies generally rely heavily on expert-annotated datasets
and unimodal approaches, limiting robustness and scalability. This study
proposes a context-based multimodal method, integrating both text and audio
data using the best-performing LPMs in each modality. By incorporating
contextual embeddings, our method improves dementia detection performance.
Additionally, motivated by the effectiveness of contextual embeddings, we
further experimented with a context-based In-Context Learning (ICL) as a
complementary technique. Results show that GPT-based embeddings, particularly
when fused with CLAP audio features, achieve an F1-score of $83.33\%$,
surpassing state-of-the-art dementia detection models. Furthermore, raw text
data outperforms expert-annotated datasets, demonstrating that LPMs can extract
meaningful linguistic and acoustic patterns without extensive manual labeling.
These findings highlight the potential for scalable, non-invasive diagnostic
tools that reduce reliance on costly annotations while maintaining high
accuracy. By integrating multimodal learning with contextual embeddings, this
work lays the foundation for future advancements in personalized dementia
detection and cognitive health research.
|
no_new_dataset
| 0.943556 |
2503.01227
|
Victor Fung
|
Shuyi Jia, Shitij Govil, Manav Ramprasad, Victor Fung
|
Pre-training Graph Neural Networks with Structural Fingerprints for
Materials Discovery
| null | null | null | null |
cond-mat.mtrl-sci cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
In recent years, pre-trained graph neural networks (GNNs) have been developed
as general models which can be effectively fine-tuned for various potential
downstream tasks in materials science, and have shown significant improvements
in accuracy and data efficiency. The most widely used pre-training methods
currently involve either supervised training to fit a general force field or
self-supervised training by denoising atomic structures equilibrium. Both
methods require datasets generated from quantum mechanical calculations, which
quickly become intractable when scaling to larger datasets. Here we propose a
novel pre-training objective which instead uses cheaply-computed structural
fingerprints as targets while maintaining comparable performance across a range
of different structural descriptors. Our experiments show this approach can act
as a general strategy for pre-training GNNs with application towards large
scale foundational models for atomistic data.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:50:23 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Jia",
"Shuyi",
""
],
[
"Govil",
"Shitij",
""
],
[
"Ramprasad",
"Manav",
""
],
[
"Fung",
"Victor",
""
]
] |
TITLE: Pre-training Graph Neural Networks with Structural Fingerprints for
Materials Discovery
ABSTRACT: In recent years, pre-trained graph neural networks (GNNs) have been developed
as general models which can be effectively fine-tuned for various potential
downstream tasks in materials science, and have shown significant improvements
in accuracy and data efficiency. The most widely used pre-training methods
currently involve either supervised training to fit a general force field or
self-supervised training by denoising atomic structures equilibrium. Both
methods require datasets generated from quantum mechanical calculations, which
quickly become intractable when scaling to larger datasets. Here we propose a
novel pre-training objective which instead uses cheaply-computed structural
fingerprints as targets while maintaining comparable performance across a range
of different structural descriptors. Our experiments show this approach can act
as a general strategy for pre-training GNNs with application towards large
scale foundational models for atomistic data.
|
no_new_dataset
| 0.951594 |
2503.01229
|
Ahmad Taha
|
Aleksandar Avdalovic and Joseph Khoury and Ahmad Taha and Elias
Bou-Harb
|
Enhancing Network Security Management in Water Systems using FM-based
Attack Attribution
| null | null | null | null |
cs.LG cs.CR cs.SY eess.SY
|
http://creativecommons.org/licenses/by/4.0/
|
Water systems are vital components of modern infrastructure, yet they are
increasingly susceptible to sophisticated cyber attacks with potentially dire
consequences on public health and safety. While state-of-the-art machine
learning techniques effectively detect anomalies, contemporary model-agnostic
attack attribution methods using LIME, SHAP, and LEMNA are deemed impractical
for large-scale, interdependent water systems. This is due to the intricate
interconnectivity and dynamic interactions that define these complex
environments. Such methods primarily emphasize individual feature importance
while falling short of addressing the crucial sensor-actuator interactions in
water systems, which limits their effectiveness in identifying root cause
attacks. To this end, we propose a novel model-agnostic Factorization Machines
(FM)-based approach that capitalizes on water system sensor-actuator
interactions to provide granular explanations and attributions for cyber
attacks. For instance, an anomaly in an actuator pump activity can be
attributed to a top root cause attack candidates, a list of water pressure
sensors, which is derived from the underlying linear and quadratic effects
captured by our approach. We validate our method using two real-world water
system specific datasets, SWaT and WADI, demonstrating its superior performance
over traditional attribution methods. In multi-feature cyber attack scenarios
involving intricate sensor-actuator interactions, our FM-based attack
attribution method effectively ranks attack root causes, achieving
approximately 20% average improvement over SHAP and LEMNA.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:52:00 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Avdalovic",
"Aleksandar",
""
],
[
"Khoury",
"Joseph",
""
],
[
"Taha",
"Ahmad",
""
],
[
"Bou-Harb",
"Elias",
""
]
] |
TITLE: Enhancing Network Security Management in Water Systems using FM-based
Attack Attribution
ABSTRACT: Water systems are vital components of modern infrastructure, yet they are
increasingly susceptible to sophisticated cyber attacks with potentially dire
consequences on public health and safety. While state-of-the-art machine
learning techniques effectively detect anomalies, contemporary model-agnostic
attack attribution methods using LIME, SHAP, and LEMNA are deemed impractical
for large-scale, interdependent water systems. This is due to the intricate
interconnectivity and dynamic interactions that define these complex
environments. Such methods primarily emphasize individual feature importance
while falling short of addressing the crucial sensor-actuator interactions in
water systems, which limits their effectiveness in identifying root cause
attacks. To this end, we propose a novel model-agnostic Factorization Machines
(FM)-based approach that capitalizes on water system sensor-actuator
interactions to provide granular explanations and attributions for cyber
attacks. For instance, an anomaly in an actuator pump activity can be
attributed to a top root cause attack candidates, a list of water pressure
sensors, which is derived from the underlying linear and quadratic effects
captured by our approach. We validate our method using two real-world water
system specific datasets, SWaT and WADI, demonstrating its superior performance
over traditional attribution methods. In multi-feature cyber attack scenarios
involving intricate sensor-actuator interactions, our FM-based attack
attribution method effectively ranks attack root causes, achieving
approximately 20% average improvement over SHAP and LEMNA.
|
no_new_dataset
| 0.948822 |
2503.01235
|
Aman Sinha
|
Timothee Mickus, Aman Sinha, Ra\'ul V\'azquez
|
Your Model is Overconfident, and Other Lies We Tell Ourselves
| null | null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
The difficulty intrinsic to a given example, rooted in its inherent
ambiguity, is a key yet often overlooked factor in evaluating neural NLP
models. We investigate the interplay and divergence among various metrics for
assessing intrinsic difficulty, including annotator dissensus, training
dynamics, and model confidence. Through a comprehensive analysis using 29
models on three datasets, we reveal that while correlations exist among these
metrics, their relationships are neither linear nor monotonic. By disentangling
these dimensions of uncertainty, we aim to refine our understanding of data
complexity and its implications for evaluating and improving NLP models.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 06:59:28 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Mickus",
"Timothee",
""
],
[
"Sinha",
"Aman",
""
],
[
"Vázquez",
"Raúl",
""
]
] |
TITLE: Your Model is Overconfident, and Other Lies We Tell Ourselves
ABSTRACT: The difficulty intrinsic to a given example, rooted in its inherent
ambiguity, is a key yet often overlooked factor in evaluating neural NLP
models. We investigate the interplay and divergence among various metrics for
assessing intrinsic difficulty, including annotator dissensus, training
dynamics, and model confidence. Through a comprehensive analysis using 29
models on three datasets, we reveal that while correlations exist among these
metrics, their relationships are neither linear nor monotonic. By disentangling
these dimensions of uncertainty, we aim to refine our understanding of data
complexity and its implications for evaluating and improving NLP models.
|
no_new_dataset
| 0.949389 |
2503.01238
|
Jensen Gao
|
Jensen Gao, Suneel Belkhale, Sudeep Dasari, Ashwin Balakrishna, Dhruv
Shah, Dorsa Sadigh
|
A Taxonomy for Evaluating Generalist Robot Policies
|
25 pages
| null | null | null |
cs.RO cs.AI cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Machine learning for robotics promises to unlock generalization to novel
tasks and environments. Guided by this promise, many recent works have focused
on scaling up robot data collection and developing larger, more expressive
policies to achieve this. But how do we measure progress towards this goal of
policy generalization in practice? Evaluating and quantifying generalization is
the Wild West of modern robotics, with each work proposing and measuring
different types of generalization in their own, often difficult to reproduce,
settings. In this work, our goal is (1) to outline the forms of generalization
we believe are important in robot manipulation in a comprehensive and
fine-grained manner, and (2) to provide reproducible guidelines for measuring
these notions of generalization. We first propose STAR-Gen, a taxonomy of
generalization for robot manipulation structured around visual, semantic, and
behavioral generalization. We discuss how our taxonomy encompasses most prior
notions of generalization in robotics. Next, we instantiate STAR-Gen with a
concrete real-world benchmark based on the widely-used Bridge V2 dataset. We
evaluate a variety of state-of-the-art models on this benchmark to demonstrate
the utility of our taxonomy in practice. Our taxonomy of generalization can
yield many interesting insights into existing models: for example, we observe
that current vision-language-action models struggle with various types of
semantic generalization, despite the promise of pre-training on internet-scale
language datasets. We believe STAR-Gen and our guidelines can improve the
dissemination and evaluation of progress towards generalization in robotics,
which we hope will guide model design and future data collection efforts. We
provide videos and demos at our website stargen-taxonomy.github.io.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 07:03:00 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Gao",
"Jensen",
""
],
[
"Belkhale",
"Suneel",
""
],
[
"Dasari",
"Sudeep",
""
],
[
"Balakrishna",
"Ashwin",
""
],
[
"Shah",
"Dhruv",
""
],
[
"Sadigh",
"Dorsa",
""
]
] |
TITLE: A Taxonomy for Evaluating Generalist Robot Policies
ABSTRACT: Machine learning for robotics promises to unlock generalization to novel
tasks and environments. Guided by this promise, many recent works have focused
on scaling up robot data collection and developing larger, more expressive
policies to achieve this. But how do we measure progress towards this goal of
policy generalization in practice? Evaluating and quantifying generalization is
the Wild West of modern robotics, with each work proposing and measuring
different types of generalization in their own, often difficult to reproduce,
settings. In this work, our goal is (1) to outline the forms of generalization
we believe are important in robot manipulation in a comprehensive and
fine-grained manner, and (2) to provide reproducible guidelines for measuring
these notions of generalization. We first propose STAR-Gen, a taxonomy of
generalization for robot manipulation structured around visual, semantic, and
behavioral generalization. We discuss how our taxonomy encompasses most prior
notions of generalization in robotics. Next, we instantiate STAR-Gen with a
concrete real-world benchmark based on the widely-used Bridge V2 dataset. We
evaluate a variety of state-of-the-art models on this benchmark to demonstrate
the utility of our taxonomy in practice. Our taxonomy of generalization can
yield many interesting insights into existing models: for example, we observe
that current vision-language-action models struggle with various types of
semantic generalization, despite the promise of pre-training on internet-scale
language datasets. We believe STAR-Gen and our guidelines can improve the
dissemination and evaluation of progress towards generalization in robotics,
which we hope will guide model design and future data collection efforts. We
provide videos and demos at our website stargen-taxonomy.github.io.
|
no_new_dataset
| 0.940408 |
2503.01254
|
Xiaolong Yu
|
Xiaolong Yu, Junqiao Zhao, Shuangfu Song, Zhongyang Zhu, Zihan Yuan,
Chen Ye, Tiantian Feng
|
Convex Hull-based Algebraic Constraint for Visual Quadric SLAM
| null | null | null | null |
cs.CV cs.RO
|
http://creativecommons.org/licenses/by/4.0/
|
Using Quadrics as the object representation has the benefits of both
generality and closed-form projection derivation between image and world
spaces. Although numerous constraints have been proposed for dual quadric
reconstruction, we found that many of them are imprecise and provide minimal
improvements to localization.After scrutinizing the existing constraints, we
introduce a concise yet more precise convex hull-based algebraic constraint for
object landmarks, which is applied to object reconstruction, frontend pose
estimation, and backend bundle adjustment.This constraint is designed to fully
leverage precise semantic segmentation, effectively mitigating mismatches
between complex-shaped object contours and dual quadrics.Experiments on public
datasets demonstrate that our approach is applicable to both monocular and
RGB-D SLAM and achieves improved object mapping and localization than existing
quadric SLAM methods. The implementation of our method is available at
https://github.com/tiev-tongji/convexhull-based-algebraic-constraint.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 07:30:07 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Yu",
"Xiaolong",
""
],
[
"Zhao",
"Junqiao",
""
],
[
"Song",
"Shuangfu",
""
],
[
"Zhu",
"Zhongyang",
""
],
[
"Yuan",
"Zihan",
""
],
[
"Ye",
"Chen",
""
],
[
"Feng",
"Tiantian",
""
]
] |
TITLE: Convex Hull-based Algebraic Constraint for Visual Quadric SLAM
ABSTRACT: Using Quadrics as the object representation has the benefits of both
generality and closed-form projection derivation between image and world
spaces. Although numerous constraints have been proposed for dual quadric
reconstruction, we found that many of them are imprecise and provide minimal
improvements to localization.After scrutinizing the existing constraints, we
introduce a concise yet more precise convex hull-based algebraic constraint for
object landmarks, which is applied to object reconstruction, frontend pose
estimation, and backend bundle adjustment.This constraint is designed to fully
leverage precise semantic segmentation, effectively mitigating mismatches
between complex-shaped object contours and dual quadrics.Experiments on public
datasets demonstrate that our approach is applicable to both monocular and
RGB-D SLAM and achieves improved object mapping and localization than existing
quadric SLAM methods. The implementation of our method is available at
https://github.com/tiev-tongji/convexhull-based-algebraic-constraint.
|
no_new_dataset
| 0.948632 |
2503.01256
|
Yuxin Wang
|
Yuxin Wang, Botian Jiang, Yiran Guo, Quan Gan, David Wipf, Xuanjing
Huang, Xipeng Qiu
|
Prior-Fitted Networks Scale to Larger Datasets When Treated as Weak
Learners
|
AISTATS 2025
| null | null | null |
cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Prior-Fitted Networks (PFNs) have recently been proposed to efficiently
perform tabular classification tasks. Although they achieve good performance on
small datasets, they encounter limitations with larger datasets. These
limitations include significant memory consumption and increased computational
complexity, primarily due to the impracticality of incorporating all training
samples as inputs within these networks. To address these challenges, we
investigate the fitting assumption for PFNs and input samples. Building on this
understanding, we propose \textit{BoostPFN} designed to enhance the performance
of these networks, especially for large-scale datasets. We also theoretically
validate the convergence of BoostPFN and our empirical results demonstrate that
the BoostPFN method can outperform standard PFNs with the same size of training
samples in large datasets and achieve a significant acceleration in training
times compared to other established baselines in the field, including
widely-used Gradient Boosting Decision Trees (GBDTs), deep learning methods and
AutoML systems. High performance is maintained for up to 50x of the
pre-training size of PFNs, substantially extending the limit of training
samples. Through this work, we address the challenges of efficiently handling
large datasets via PFN-based models, paving the way for faster and more
effective tabular data classification training and prediction process. Code is
available at Github.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 07:31:40 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Wang",
"Yuxin",
""
],
[
"Jiang",
"Botian",
""
],
[
"Guo",
"Yiran",
""
],
[
"Gan",
"Quan",
""
],
[
"Wipf",
"David",
""
],
[
"Huang",
"Xuanjing",
""
],
[
"Qiu",
"Xipeng",
""
]
] |
TITLE: Prior-Fitted Networks Scale to Larger Datasets When Treated as Weak
Learners
ABSTRACT: Prior-Fitted Networks (PFNs) have recently been proposed to efficiently
perform tabular classification tasks. Although they achieve good performance on
small datasets, they encounter limitations with larger datasets. These
limitations include significant memory consumption and increased computational
complexity, primarily due to the impracticality of incorporating all training
samples as inputs within these networks. To address these challenges, we
investigate the fitting assumption for PFNs and input samples. Building on this
understanding, we propose \textit{BoostPFN} designed to enhance the performance
of these networks, especially for large-scale datasets. We also theoretically
validate the convergence of BoostPFN and our empirical results demonstrate that
the BoostPFN method can outperform standard PFNs with the same size of training
samples in large datasets and achieve a significant acceleration in training
times compared to other established baselines in the field, including
widely-used Gradient Boosting Decision Trees (GBDTs), deep learning methods and
AutoML systems. High performance is maintained for up to 50x of the
pre-training size of PFNs, substantially extending the limit of training
samples. Through this work, we address the challenges of efficiently handling
large datasets via PFN-based models, paving the way for faster and more
effective tabular data classification training and prediction process. Code is
available at Github.
|
no_new_dataset
| 0.944638 |
2503.01257
|
Xuan Zhu
|
Xuan Zhu, Jijun Xiang, Xianqi Wang, Longliang Liu, Yu Wang, Hong
Zhang, Fei Guo, Xin Yang
|
SVDC: Consistent Direct Time-of-Flight Video Depth Completion with
Frequency Selective Fusion
|
Accepted by CVPR 2025
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Lightweight direct Time-of-Flight (dToF) sensors are ideal for 3D sensing on
mobile devices. However, due to the manufacturing constraints of compact
devices and the inherent physical principles of imaging, dToF depth maps are
sparse and noisy. In this paper, we propose a novel video depth completion
method, called SVDC, by fusing the sparse dToF data with the corresponding RGB
guidance. Our method employs a multi-frame fusion scheme to mitigate the
spatial ambiguity resulting from the sparse dToF imaging. Misalignment between
consecutive frames during multi-frame fusion could cause blending between
object edges and the background, which results in a loss of detail. To address
this, we introduce an adaptive frequency selective fusion (AFSF) module, which
automatically selects convolution kernel sizes to fuse multi-frame features.
Our AFSF utilizes a channel-spatial enhancement attention (CSEA) module to
enhance features and generates an attention map as fusion weights. The AFSF
ensures edge detail recovery while suppressing high-frequency noise in smooth
regions. To further enhance temporal consistency, We propose a cross-window
consistency loss to ensure consistent predictions across different windows,
effectively reducing flickering. Our proposed SVDC achieves optimal accuracy
and consistency on the TartanAir and Dynamic Replica datasets. Code is
available at https://github.com/Lan1eve/SVDC.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 07:32:25 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Zhu",
"Xuan",
""
],
[
"Xiang",
"Jijun",
""
],
[
"Wang",
"Xianqi",
""
],
[
"Liu",
"Longliang",
""
],
[
"Wang",
"Yu",
""
],
[
"Zhang",
"Hong",
""
],
[
"Guo",
"Fei",
""
],
[
"Yang",
"Xin",
""
]
] |
TITLE: SVDC: Consistent Direct Time-of-Flight Video Depth Completion with
Frequency Selective Fusion
ABSTRACT: Lightweight direct Time-of-Flight (dToF) sensors are ideal for 3D sensing on
mobile devices. However, due to the manufacturing constraints of compact
devices and the inherent physical principles of imaging, dToF depth maps are
sparse and noisy. In this paper, we propose a novel video depth completion
method, called SVDC, by fusing the sparse dToF data with the corresponding RGB
guidance. Our method employs a multi-frame fusion scheme to mitigate the
spatial ambiguity resulting from the sparse dToF imaging. Misalignment between
consecutive frames during multi-frame fusion could cause blending between
object edges and the background, which results in a loss of detail. To address
this, we introduce an adaptive frequency selective fusion (AFSF) module, which
automatically selects convolution kernel sizes to fuse multi-frame features.
Our AFSF utilizes a channel-spatial enhancement attention (CSEA) module to
enhance features and generates an attention map as fusion weights. The AFSF
ensures edge detail recovery while suppressing high-frequency noise in smooth
regions. To further enhance temporal consistency, We propose a cross-window
consistency loss to ensure consistent predictions across different windows,
effectively reducing flickering. Our proposed SVDC achieves optimal accuracy
and consistency on the TartanAir and Dynamic Replica datasets. Code is
available at https://github.com/Lan1eve/SVDC.
|
no_new_dataset
| 0.948058 |
2503.01260
|
Yuhan Jing
|
Yuhan Jing, Jingyu Wang, Lei Zhang, Haifeng Sun, Bo He, Zirui Zhuang,
Chengsen Wang, Qi Qi, Jianxin Liao
|
OIPR: Evaluation for Time-series Anomaly Detection Inspired by Operator
Interest
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
With the growing adoption of time-series anomaly detection (TAD) technology,
numerous studies have employed deep learning-based detectors for analyzing
time-series data in the fields of Internet services, industrial systems, and
sensors. The selection and optimization of anomaly detectors strongly rely on
the availability of an effective performance evaluation method for TAD. Since
anomalies in time-series data often manifest as a sequence of points,
conventional metrics that solely consider the detection of individual point are
inadequate. Existing evaluation methods for TAD typically employ point-based or
event-based metrics to capture the temporal context. However, point-based
metrics tend to overestimate detectors that excel only in detecting long
anomalies, while event-based metrics are susceptible to being misled by
fragmented detection results. To address these limitations, we propose OIPR, a
novel set of TAD evaluation metrics. It models the process of operators
receiving detector alarms and handling faults, utilizing area under the
operator interest curve to evaluate the performance of TAD algorithms.
Furthermore, we build a special scenario dataset to compare the characteristics
of different evaluation methods. Through experiments conducted on the special
scenario dataset and five real-world datasets, we demonstrate the remarkable
performance of OIPR in extreme and complex scenarios. It achieves a balance
between point and event perspectives, overcoming their primary limitations and
offering applicability to broader situations.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 07:37:24 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Jing",
"Yuhan",
""
],
[
"Wang",
"Jingyu",
""
],
[
"Zhang",
"Lei",
""
],
[
"Sun",
"Haifeng",
""
],
[
"He",
"Bo",
""
],
[
"Zhuang",
"Zirui",
""
],
[
"Wang",
"Chengsen",
""
],
[
"Qi",
"Qi",
""
],
[
"Liao",
"Jianxin",
""
]
] |
TITLE: OIPR: Evaluation for Time-series Anomaly Detection Inspired by Operator
Interest
ABSTRACT: With the growing adoption of time-series anomaly detection (TAD) technology,
numerous studies have employed deep learning-based detectors for analyzing
time-series data in the fields of Internet services, industrial systems, and
sensors. The selection and optimization of anomaly detectors strongly rely on
the availability of an effective performance evaluation method for TAD. Since
anomalies in time-series data often manifest as a sequence of points,
conventional metrics that solely consider the detection of individual point are
inadequate. Existing evaluation methods for TAD typically employ point-based or
event-based metrics to capture the temporal context. However, point-based
metrics tend to overestimate detectors that excel only in detecting long
anomalies, while event-based metrics are susceptible to being misled by
fragmented detection results. To address these limitations, we propose OIPR, a
novel set of TAD evaluation metrics. It models the process of operators
receiving detector alarms and handling faults, utilizing area under the
operator interest curve to evaluate the performance of TAD algorithms.
Furthermore, we build a special scenario dataset to compare the characteristics
of different evaluation methods. Through experiments conducted on the special
scenario dataset and five real-world datasets, we demonstrate the remarkable
performance of OIPR in extreme and complex scenarios. It achieves a balance
between point and event perspectives, overcoming their primary limitations and
offering applicability to broader situations.
|
new_dataset
| 0.967318 |
2503.01266
|
Birger Moell
|
Birger Moell, Fredrik Sand Aronsson
|
Voice Cloning for Dysarthric Speech Synthesis: Addressing Data Scarcity
in Speech-Language Pathology
| null | null | null | null |
cs.SD cs.AI eess.AS
|
http://creativecommons.org/licenses/by/4.0/
|
This study explores voice cloning to generate synthetic speech replicating
the unique patterns of individuals with dysarthria. Using the TORGO dataset, we
address data scarcity and privacy challenges in speech-language pathology. Our
contributions include demonstrating that voice cloning preserves dysarthric
speech characteristics, analyzing differences between real and synthetic data,
and discussing implications for diagnostics, rehabilitation, and communication.
We cloned voices from dysarthric and control speakers using a commercial
platform, ensuring gender-matched synthetic voices. A licensed speech-language
pathologist (SLP) evaluated a subset for dysarthria, speaker gender, and
synthetic indicators. The SLP correctly identified dysarthria in all cases and
speaker gender in 95% but misclassified 30% of synthetic samples as real,
indicating high realism. Our results suggest synthetic speech effectively
captures disordered characteristics and that voice cloning has advanced to
produce high-quality data resembling real speech, even to trained
professionals. This has critical implications for healthcare, where synthetic
data can mitigate data scarcity, protect privacy, and enhance AI-driven
diagnostics. By enabling the creation of diverse, high-quality speech datasets,
voice cloning can improve generalizable models, personalize therapy, and
advance assistive technologies for dysarthria.
We publicly release our synthetic dataset to foster further research and
collaboration, aiming to develop robust models that improve patient outcomes in
speech-language pathology.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 07:44:49 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Moell",
"Birger",
""
],
[
"Aronsson",
"Fredrik Sand",
""
]
] |
TITLE: Voice Cloning for Dysarthric Speech Synthesis: Addressing Data Scarcity
in Speech-Language Pathology
ABSTRACT: This study explores voice cloning to generate synthetic speech replicating
the unique patterns of individuals with dysarthria. Using the TORGO dataset, we
address data scarcity and privacy challenges in speech-language pathology. Our
contributions include demonstrating that voice cloning preserves dysarthric
speech characteristics, analyzing differences between real and synthetic data,
and discussing implications for diagnostics, rehabilitation, and communication.
We cloned voices from dysarthric and control speakers using a commercial
platform, ensuring gender-matched synthetic voices. A licensed speech-language
pathologist (SLP) evaluated a subset for dysarthria, speaker gender, and
synthetic indicators. The SLP correctly identified dysarthria in all cases and
speaker gender in 95% but misclassified 30% of synthetic samples as real,
indicating high realism. Our results suggest synthetic speech effectively
captures disordered characteristics and that voice cloning has advanced to
produce high-quality data resembling real speech, even to trained
professionals. This has critical implications for healthcare, where synthetic
data can mitigate data scarcity, protect privacy, and enhance AI-driven
diagnostics. By enabling the creation of diverse, high-quality speech datasets,
voice cloning can improve generalizable models, personalize therapy, and
advance assistive technologies for dysarthria.
We publicly release our synthetic dataset to foster further research and
collaboration, aiming to develop robust models that improve patient outcomes in
speech-language pathology.
|
new_dataset
| 0.952926 |
2503.01273
|
Yuxuan Chen
|
Yuxuan Chen, Long Zhang, Xu Zhu, Hua Zhou, Zhuyin Ren
|
OptMetaOpenFOAM: Large Language Model Driven Chain of Thought for
Sensitivity Analysis and Parameter Optimization based on CFD
|
26 pages,11 figures
| null | null | null |
cs.AI physics.flu-dyn
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Merging natural language interfaces with computational fluid dynamics (CFD)
workflows presents transformative opportunities for both industry and research.
In this study, we introduce OptMetaOpenFOAM - a novel framework that bridges
MetaOpenFOAM with external analysis and optimization tool libraries through a
large language model (LLM)-driven chain-of-thought (COT) methodology. By
automating complex CFD tasks via natural language inputs, the framework
empowers non-expert users to perform sensitivity analyses and parameter
optimizations with markedly improved efficiency. The test dataset comprises 11
distinct CFD analysis or optimization tasks, including a baseline simulation
task derived from an OpenFOAM tutorial covering fluid dynamics, combustion, and
heat transfer. Results confirm that OptMetaOpenFOAM can accurately interpret
user requirements expressed in natural language and effectively invoke external
tool libraries alongside MetaOpenFOAM to complete the tasks. Furthermore,
validation on a non-OpenFOAM tutorial case - namely, a hydrogen combustion
chamber - demonstrates that a mere 200-character natural language input can
trigger a sequence of simulation, postprocessing, analysis, and optimization
tasks spanning over 2,000 lines of code. These findings underscore the
transformative potential of LLM-driven COT methodologies in linking external
tool for advanced analysis and optimization, positioning OptMetaOpenFOAM as an
effective tool that streamlines CFD simulations and enhances their convenience
and efficiency for both industrial and research applications. Code is available
at https://github.com/Terry-cyx/MetaOpenFOAM.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 07:55:43 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Chen",
"Yuxuan",
""
],
[
"Zhang",
"Long",
""
],
[
"Zhu",
"Xu",
""
],
[
"Zhou",
"Hua",
""
],
[
"Ren",
"Zhuyin",
""
]
] |
TITLE: OptMetaOpenFOAM: Large Language Model Driven Chain of Thought for
Sensitivity Analysis and Parameter Optimization based on CFD
ABSTRACT: Merging natural language interfaces with computational fluid dynamics (CFD)
workflows presents transformative opportunities for both industry and research.
In this study, we introduce OptMetaOpenFOAM - a novel framework that bridges
MetaOpenFOAM with external analysis and optimization tool libraries through a
large language model (LLM)-driven chain-of-thought (COT) methodology. By
automating complex CFD tasks via natural language inputs, the framework
empowers non-expert users to perform sensitivity analyses and parameter
optimizations with markedly improved efficiency. The test dataset comprises 11
distinct CFD analysis or optimization tasks, including a baseline simulation
task derived from an OpenFOAM tutorial covering fluid dynamics, combustion, and
heat transfer. Results confirm that OptMetaOpenFOAM can accurately interpret
user requirements expressed in natural language and effectively invoke external
tool libraries alongside MetaOpenFOAM to complete the tasks. Furthermore,
validation on a non-OpenFOAM tutorial case - namely, a hydrogen combustion
chamber - demonstrates that a mere 200-character natural language input can
trigger a sequence of simulation, postprocessing, analysis, and optimization
tasks spanning over 2,000 lines of code. These findings underscore the
transformative potential of LLM-driven COT methodologies in linking external
tool for advanced analysis and optimization, positioning OptMetaOpenFOAM as an
effective tool that streamlines CFD simulations and enhances their convenience
and efficiency for both industrial and research applications. Code is available
at https://github.com/Terry-cyx/MetaOpenFOAM.
|
no_new_dataset
| 0.942188 |
2503.01287
|
Yogesh Verma
|
Yogesh Verma, Ayush Bharti and Vikas Garg
|
Robust Simulation-Based Inference under Missing Data via Neural
Processes
|
Accepted at ICLR 2025
| null | null | null |
cs.LG cs.AI stat.ML
|
http://creativecommons.org/licenses/by-sa/4.0/
|
Simulation-based inference (SBI) methods typically require fully observed
data to infer parameters of models with intractable likelihood functions.
However, datasets often contain missing values due to incomplete observations,
data corruptions (common in astrophysics), or instrument limitations (e.g., in
high-energy physics applications). In such scenarios, missing data must be
imputed before applying any SBI method. We formalize the problem of missing
data in SBI and demonstrate that naive imputation methods can introduce bias in
the estimation of SBI posterior. We also introduce a novel amortized method
that addresses this issue by jointly learning the imputation model and the
inference network within a neural posterior estimation (NPE) framework.
Extensive empirical results on SBI benchmarks show that our approach provides
robust inference outcomes compared to standard baselines for varying levels of
missing data. Moreover, we demonstrate the merits of our imputation model on
two real-world bioactivity datasets (Adrenergic and Kinase assays). Code is
available at https://github.com/Aalto-QuML/RISE.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 08:22:01 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Verma",
"Yogesh",
""
],
[
"Bharti",
"Ayush",
""
],
[
"Garg",
"Vikas",
""
]
] |
TITLE: Robust Simulation-Based Inference under Missing Data via Neural
Processes
ABSTRACT: Simulation-based inference (SBI) methods typically require fully observed
data to infer parameters of models with intractable likelihood functions.
However, datasets often contain missing values due to incomplete observations,
data corruptions (common in astrophysics), or instrument limitations (e.g., in
high-energy physics applications). In such scenarios, missing data must be
imputed before applying any SBI method. We formalize the problem of missing
data in SBI and demonstrate that naive imputation methods can introduce bias in
the estimation of SBI posterior. We also introduce a novel amortized method
that addresses this issue by jointly learning the imputation model and the
inference network within a neural posterior estimation (NPE) framework.
Extensive empirical results on SBI benchmarks show that our approach provides
robust inference outcomes compared to standard baselines for varying levels of
missing data. Moreover, we demonstrate the merits of our imputation model on
two real-world bioactivity datasets (Adrenergic and Kinase assays). Code is
available at https://github.com/Aalto-QuML/RISE.
|
no_new_dataset
| 0.945951 |
2503.01288
|
Chong Wang
|
Chong Wang, Lanqing Guo, Zixuan Fu, Siyuan Yang, Hao Cheng, Alex C.
Kot, Bihan Wen
|
Reconciling Stochastic and Deterministic Strategies for Zero-shot Image
Restoration using Diffusion Model in Dual
|
Accepted to CVPR 2025
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Plug-and-play (PnP) methods offer an iterative strategy for solving image
restoration (IR) problems in a zero-shot manner, using a learned
\textit{discriminative denoiser} as the implicit prior. More recently, a
sampling-based variant of this approach, which utilizes a pre-trained
\textit{generative diffusion model}, has gained great popularity for solving IR
problems through stochastic sampling. The IR results using PnP with a
pre-trained diffusion model demonstrate distinct advantages compared to those
using discriminative denoisers, \ie improved perceptual quality while
sacrificing the data fidelity. The unsatisfactory results are due to the lack
of integration of these strategies in the IR tasks. In this work, we propose a
novel zero-shot IR scheme, dubbed Reconciling Diffusion Model in Dual (RDMD),
which leverages only a \textbf{single} pre-trained diffusion model to construct
\textbf{two} complementary regularizers. Specifically, the diffusion model in
RDMD will iteratively perform deterministic denoising and stochastic sampling,
aiming to achieve high-fidelity image restoration with appealing perceptual
quality. RDMD also allows users to customize the distortion-perception tradeoff
with a single hyperparameter, enhancing the adaptability of the restoration
process in different practical scenarios. Extensive experiments on several IR
tasks demonstrate that our proposed method could achieve superior results
compared to existing approaches on both the FFHQ and ImageNet datasets.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 08:25:22 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Wang",
"Chong",
""
],
[
"Guo",
"Lanqing",
""
],
[
"Fu",
"Zixuan",
""
],
[
"Yang",
"Siyuan",
""
],
[
"Cheng",
"Hao",
""
],
[
"Kot",
"Alex C.",
""
],
[
"Wen",
"Bihan",
""
]
] |
TITLE: Reconciling Stochastic and Deterministic Strategies for Zero-shot Image
Restoration using Diffusion Model in Dual
ABSTRACT: Plug-and-play (PnP) methods offer an iterative strategy for solving image
restoration (IR) problems in a zero-shot manner, using a learned
\textit{discriminative denoiser} as the implicit prior. More recently, a
sampling-based variant of this approach, which utilizes a pre-trained
\textit{generative diffusion model}, has gained great popularity for solving IR
problems through stochastic sampling. The IR results using PnP with a
pre-trained diffusion model demonstrate distinct advantages compared to those
using discriminative denoisers, \ie improved perceptual quality while
sacrificing the data fidelity. The unsatisfactory results are due to the lack
of integration of these strategies in the IR tasks. In this work, we propose a
novel zero-shot IR scheme, dubbed Reconciling Diffusion Model in Dual (RDMD),
which leverages only a \textbf{single} pre-trained diffusion model to construct
\textbf{two} complementary regularizers. Specifically, the diffusion model in
RDMD will iteratively perform deterministic denoising and stochastic sampling,
aiming to achieve high-fidelity image restoration with appealing perceptual
quality. RDMD also allows users to customize the distortion-perception tradeoff
with a single hyperparameter, enhancing the adaptability of the restoration
process in different practical scenarios. Extensive experiments on several IR
tasks demonstrate that our proposed method could achieve superior results
compared to existing approaches on both the FFHQ and ImageNet datasets.
|
no_new_dataset
| 0.948251 |
2503.01290
|
Andreas Sauter
|
Andreas Sauter and Saber Salehkaleybar and Aske Plaat and Erman Acar
|
ACTIVA: Amortized Causal Effect Estimation without Graphs via
Transformer-based Variational Autoencoder
| null | null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Predicting the distribution of outcomes under hypothetical interventions is
crucial in domains like healthcare, economics, and policy-making. Current
methods often rely on strong assumptions, such as known causal graphs or
parametric models, and lack amortization across problem instances, limiting
their practicality. We propose a novel transformer-based conditional
variational autoencoder architecture, named ACTIVA, that extends causal
transformer encoders to predict causal effects as mixtures of Gaussians. Our
method requires no causal graph and predicts interventional distributions given
only observational data and a queried intervention. By amortizing over many
simulated instances, it enables zero-shot generalization to novel datasets
without retraining. Experiments demonstrate accurate predictions for synthetic
and semi-synthetic data, showcasing the effectiveness of our graph-free,
amortized causal inference approach.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 08:28:25 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Sauter",
"Andreas",
""
],
[
"Salehkaleybar",
"Saber",
""
],
[
"Plaat",
"Aske",
""
],
[
"Acar",
"Erman",
""
]
] |
TITLE: ACTIVA: Amortized Causal Effect Estimation without Graphs via
Transformer-based Variational Autoencoder
ABSTRACT: Predicting the distribution of outcomes under hypothetical interventions is
crucial in domains like healthcare, economics, and policy-making. Current
methods often rely on strong assumptions, such as known causal graphs or
parametric models, and lack amortization across problem instances, limiting
their practicality. We propose a novel transformer-based conditional
variational autoencoder architecture, named ACTIVA, that extends causal
transformer encoders to predict causal effects as mixtures of Gaussians. Our
method requires no causal graph and predicts interventional distributions given
only observational data and a queried intervention. By amortizing over many
simulated instances, it enables zero-shot generalization to novel datasets
without retraining. Experiments demonstrate accurate predictions for synthetic
and semi-synthetic data, showcasing the effectiveness of our graph-free,
amortized causal inference approach.
|
no_new_dataset
| 0.947186 |
2503.01291
|
Peishan Cong
|
Peishan Cong, Ziyi Wang, Yuexin Ma, Xiangyu Yue
|
SemGeoMo: Dynamic Contextual Human Motion Generation with Semantic and
Geometric Guidance
|
accepted by CVPR 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Generating reasonable and high-quality human interactive motions in a given
dynamic environment is crucial for understanding, modeling, transferring, and
applying human behaviors to both virtual and physical robots. In this paper, we
introduce an effective method, SemGeoMo, for dynamic contextual human motion
generation, which fully leverages the text-affordance-joint multi-level
semantic and geometric guidance in the generation process, improving the
semantic rationality and geometric correctness of generative motions. Our
method achieves state-of-the-art performance on three datasets and demonstrates
superior generalization capability for diverse interaction scenarios. The
project page and code can be found at
https://4dvlab.github.io/project_page/semgeomo/.
|
[
{
"version": "v1",
"created": "Mon, 3 Mar 2025 08:28:40 GMT"
}
] | 2025-03-04T00:00:00 |
[
[
"Cong",
"Peishan",
""
],
[
"Wang",
"Ziyi",
""
],
[
"Ma",
"Yuexin",
""
],
[
"Yue",
"Xiangyu",
""
]
] |
TITLE: SemGeoMo: Dynamic Contextual Human Motion Generation with Semantic and
Geometric Guidance
ABSTRACT: Generating reasonable and high-quality human interactive motions in a given
dynamic environment is crucial for understanding, modeling, transferring, and
applying human behaviors to both virtual and physical robots. In this paper, we
introduce an effective method, SemGeoMo, for dynamic contextual human motion
generation, which fully leverages the text-affordance-joint multi-level
semantic and geometric guidance in the generation process, improving the
semantic rationality and geometric correctness of generative motions. Our
method achieves state-of-the-art performance on three datasets and demonstrates
superior generalization capability for diverse interaction scenarios. The
project page and code can be found at
https://4dvlab.github.io/project_page/semgeomo/.
|
no_new_dataset
| 0.952882 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.