Search is not available for this dataset
id
string
submitter
string
authors
string
title
string
comments
string
journal-ref
string
doi
string
report-no
string
categories
string
license
string
abstract
string
versions
list
update_date
timestamp[s]
authors_parsed
list
prompt
string
2504.06962
Thomas Kerdreux
Thomas Kerdreux and Alexandre Tuel and Quentin Febvre and Alexis Mouche and Bertrand Chapron
Efficient Self-Supervised Learning for Earth Observation via Dynamic Dataset Curation
Accepted at CVPR Workshop : The First Workshop on Foundation and Large Vision Models in Remote Sensing
null
null
null
cs.CV cs.AI
http://creativecommons.org/licenses/by/4.0/
Self-supervised learning (SSL) has enabled the development of vision foundation models for Earth Observation (EO), demonstrating strong transferability across diverse remote sensing tasks. While prior work has focused on network architectures and training strategies, the role of dataset curation, especially in balanc...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 15:13:26 GMT" } ]
2025-04-10T00:00:00
[ [ "Kerdreux", "Thomas", "" ], [ "Tuel", "Alexandre", "" ], [ "Febvre", "Quentin", "" ], [ "Mouche", "Alexis", "" ], [ "Chapron", "Bertrand", "" ] ]
TITLE: Efficient Self-Supervised Learning for Earth Observation via Dynamic Dataset Curation ABSTRACT: Self-supervised learning (SSL) has enabled the development of vision foundation models for Earth Observation (EO), demonstrating strong transferability across diverse remote sensing tasks. While prior work has f...
2504.06963
Vladimir Bataev
Vladimir Bataev
RNN-Transducer-based Losses for Speech Recognition on Noisy Targets
Final Project Report, Bachelor's Degree in Computer Science, University of London, March 2024
null
null
null
eess.AS cs.AI cs.CL cs.LG cs.SD
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Training speech recognition systems on noisy transcripts is a significant challenge in industrial pipelines, where datasets are enormous and ensuring accurate transcription for every instance is difficult. In this work, we introduce novel loss functions to mitigate the impact of transcription errors in RNN-Transducer...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 15:18:29 GMT" } ]
2025-04-10T00:00:00
[ [ "Bataev", "Vladimir", "" ] ]
TITLE: RNN-Transducer-based Losses for Speech Recognition on Noisy Targets ABSTRACT: Training speech recognition systems on noisy transcripts is a significant challenge in industrial pipelines, where datasets are enormous and ensuring accurate transcription for every instance is difficult. In this work, we introduc...
2504.06965
Qingsong Yan
Teng Xiao, Qi Hu, Qingsong Yan, Wei Liu, Zhiwei Ye, Fei Deng
A Deep Single Image Rectification Approach for Pan-Tilt-Zoom Cameras
Accepted to ICME 2025
null
null
null
cs.CV
http://creativecommons.org/licenses/by/4.0/
Pan-Tilt-Zoom (PTZ) cameras with wide-angle lenses are widely used in surveillance but often require image rectification due to their inherent nonlinear distortions. Current deep learning approaches typically struggle to maintain fine-grained geometric details, resulting in inaccurate rectification. This paper presen...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 15:19:38 GMT" } ]
2025-04-10T00:00:00
[ [ "Xiao", "Teng", "" ], [ "Hu", "Qi", "" ], [ "Yan", "Qingsong", "" ], [ "Liu", "Wei", "" ], [ "Ye", "Zhiwei", "" ], [ "Deng", "Fei", "" ] ]
TITLE: A Deep Single Image Rectification Approach for Pan-Tilt-Zoom Cameras ABSTRACT: Pan-Tilt-Zoom (PTZ) cameras with wide-angle lenses are widely used in surveillance but often require image rectification due to their inherent nonlinear distortions. Current deep learning approaches typically struggle to maintain ...
2504.06969
Lilian Ngweta
Lilian Ngweta, Kiran Kate, Jason Tsay, Yara Rizk
Towards LLMs Robustness to Changes in Prompt Format Styles
NAACL Student Research Workshop (SRW) 2025
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Large language models (LLMs) have gained popularity in recent years for their utility in various applications. However, they are sensitive to non-semantic changes in prompt formats, where small changes in the prompt format can lead to significant performance fluctuations. In the literature, this problem is commonly r...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 15:26:00 GMT" } ]
2025-04-10T00:00:00
[ [ "Ngweta", "Lilian", "" ], [ "Kate", "Kiran", "" ], [ "Tsay", "Jason", "" ], [ "Rizk", "Yara", "" ] ]
TITLE: Towards LLMs Robustness to Changes in Prompt Format Styles ABSTRACT: Large language models (LLMs) have gained popularity in recent years for their utility in various applications. However, they are sensitive to non-semantic changes in prompt formats, where small changes in the prompt format can lead to signi...
2504.06982
Yuhang Yang
Yuhang Yang, Fengqi Liu, Yixing Lu, Qin Zhao, Pingyu Wu, Wei Zhai, Ran Yi, Yang Cao, Lizhuang Ma, Zheng-Jun Zha, Junting Dong
SIGMAN:Scaling 3D Human Gaussian Generation with Millions of Assets
project page:https://yyvhang.github.io/SIGMAN_3D/
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
3D human digitization has long been a highly pursued yet challenging task. Existing methods aim to generate high-quality 3D digital humans from single or multiple views, but remain primarily constrained by current paradigms and the scarcity of 3D human assets. Specifically, recent approaches fall into several paradig...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 15:38:18 GMT" } ]
2025-04-10T00:00:00
[ [ "Yang", "Yuhang", "" ], [ "Liu", "Fengqi", "" ], [ "Lu", "Yixing", "" ], [ "Zhao", "Qin", "" ], [ "Wu", "Pingyu", "" ], [ "Zhai", "Wei", "" ], [ "Yi", "Ran", "" ], [ "Cao", "Yang", "" ], ...
TITLE: SIGMAN:Scaling 3D Human Gaussian Generation with Millions of Assets ABSTRACT: 3D human digitization has long been a highly pursued yet challenging task. Existing methods aim to generate high-quality 3D digital humans from single or multiple views, but remain primarily constrained by current paradigms and the...
2504.06991
Ghurumuruhan Ganesan
Ghurumuruhan Ganesan
Dissimilar Batch Decompositions of Random Datasets
Accepted for publication in Sankhya A
null
null
null
cs.LG math.PR stat.ML
http://creativecommons.org/licenses/by/4.0/
For better learning, large datasets are often split into small batches and fed sequentially to the predictive model. In this paper, we study such batch decompositions from a probabilistic perspective. We assume that data points (possibly corrupted) are drawn independently from a given space and define a concept of si...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 15:58:06 GMT" } ]
2025-04-10T00:00:00
[ [ "Ganesan", "Ghurumuruhan", "" ] ]
TITLE: Dissimilar Batch Decompositions of Random Datasets ABSTRACT: For better learning, large datasets are often split into small batches and fed sequentially to the predictive model. In this paper, we study such batch decompositions from a probabilistic perspective. We assume that data points (possibly corrupted)...
2504.06997
Mingliang Pan
Mingliang Pan, Chenxu Li, Yuanzhe Zhang, Alan Mollins, Quan Wang, Ahmet T. Erdogan, Yuanyuan Hua, Zhenya Zang, Neil Finlayson, Robert K. Henderson, David Day-Uei Li
Cerebral blood flow monitoring using a deep learning implementation of the two-layer DCS analytical model with a 512 512 SPAD array
23 pages, 11 figures
null
null
null
physics.med-ph physics.bio-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Diffuse correlation spectroscopy (DCS) analyzes the autocorrelation function of photons scattered by red blood cells, enabling non-invasive, continuous measurement of deep tissue blood flow at the bedside. Multi-layer DCS models (two- and three-layer) enhance cerebral blood flow index (CBFi) sensitivity and mitigate ...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 16:09:34 GMT" } ]
2025-04-10T00:00:00
[ [ "Pan", "Mingliang", "" ], [ "Li", "Chenxu", "" ], [ "Zhang", "Yuanzhe", "" ], [ "Mollins", "Alan", "" ], [ "Wang", "Quan", "" ], [ "Erdogan", "Ahmet T.", "" ], [ "Hua", "Yuanyuan", "" ], [ "Zang", ...
TITLE: Cerebral blood flow monitoring using a deep learning implementation of the two-layer DCS analytical model with a 512 512 SPAD array ABSTRACT: Diffuse correlation spectroscopy (DCS) analyzes the autocorrelation function of photons scattered by red blood cells, enabling non-invasive, continuous measurement o...
2504.07002
Yuan Xiao
Yuan Xiao, Yuchen Chen, Shiqing Ma, Haocheng Huang, Chunrong Fang, Yanwei Chen, Weisong Sun, Yunfeng Zhu, Xiaofang Zhang, Zhenyu Chen
DeCoMa: Detecting and Purifying Code Dataset Watermarks through Dual Channel Code Abstraction
Accepted to ISSTA 2025. Code is available at https://github.com/xiaoyuanpigo/DeCoMa
null
10.1145/3728952
null
cs.CR cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Watermarking is a technique to help identify the source of data points, which can be used to help prevent the misuse of protected datasets. Existing methods on code watermarking, leveraging the idea from the backdoor research, embed stealthy triggers as watermarks.Despite their high resilience against dilution attack...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 16:19:11 GMT" } ]
2025-04-10T00:00:00
[ [ "Xiao", "Yuan", "" ], [ "Chen", "Yuchen", "" ], [ "Ma", "Shiqing", "" ], [ "Huang", "Haocheng", "" ], [ "Fang", "Chunrong", "" ], [ "Chen", "Yanwei", "" ], [ "Sun", "Weisong", "" ], [ "Zhu", "Yu...
TITLE: DeCoMa: Detecting and Purifying Code Dataset Watermarks through Dual Channel Code Abstraction ABSTRACT: Watermarking is a technique to help identify the source of data points, which can be used to help prevent the misuse of protected datasets. Existing methods on code watermarking, leveraging the idea from...
2504.07017
Yusuf Guven
Yusuf Guven, Tufan Kumbasar
Adapting GT2-FLS for Uncertainty Quantification: A Blueprint Calibration Strategy
in IEEE International Conference on Fuzzy Systems, 2025
null
null
null
cs.LG
http://creativecommons.org/licenses/by-nc-sa/4.0/
Uncertainty Quantification (UQ) is crucial for deploying reliable Deep Learning (DL) models in high-stakes applications. Recently, General Type-2 Fuzzy Logic Systems (GT2-FLSs) have been proven to be effective for UQ, offering Prediction Intervals (PIs) to capture uncertainty. However, existing methods often struggle...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 16:32:43 GMT" } ]
2025-04-10T00:00:00
[ [ "Guven", "Yusuf", "" ], [ "Kumbasar", "Tufan", "" ] ]
TITLE: Adapting GT2-FLS for Uncertainty Quantification: A Blueprint Calibration Strategy ABSTRACT: Uncertainty Quantification (UQ) is crucial for deploying reliable Deep Learning (DL) models in high-stakes applications. Recently, General Type-2 Fuzzy Logic Systems (GT2-FLSs) have been proven to be effective for U...
2504.07025
Bojian Wu
Bojian Wu, Yifan Peng, Ruizhen Hu, Xiaowei Zhou
Glossy Object Reconstruction with Cost-effective Polarized Acquisition
Accepted to CVPR 2025 as highlight
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The challenge of image-based 3D reconstruction for glossy objects lies in separating diffuse and specular components on glossy surfaces from captured images, a task complicated by the ambiguity in discerning lighting conditions and material properties using RGB data alone. While state-of-the-art methods rely on tailo...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 16:38:51 GMT" } ]
2025-04-10T00:00:00
[ [ "Wu", "Bojian", "" ], [ "Peng", "Yifan", "" ], [ "Hu", "Ruizhen", "" ], [ "Zhou", "Xiaowei", "" ] ]
TITLE: Glossy Object Reconstruction with Cost-effective Polarized Acquisition ABSTRACT: The challenge of image-based 3D reconstruction for glossy objects lies in separating diffuse and specular components on glossy surfaces from captured images, a task complicated by the ambiguity in discerning lighting conditions ...
2504.07031
Pawel Pukowski
Pawel Pukowski and Venet Osmani
Identifying Key Challenges of Hardness-Based Resampling
Submitted to IEEE TPAMI
null
null
null
cs.LG
http://creativecommons.org/licenses/by/4.0/
Performance gap across classes remains a persistent challenge in machine learning, often attributed to variations in class hardness. One way to quantify class hardness is through sample complexity - the minimum number of samples required to effectively learn a given class. Sample complexity theory suggests that class...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 16:45:57 GMT" } ]
2025-04-10T00:00:00
[ [ "Pukowski", "Pawel", "" ], [ "Osmani", "Venet", "" ] ]
TITLE: Identifying Key Challenges of Hardness-Based Resampling ABSTRACT: Performance gap across classes remains a persistent challenge in machine learning, often attributed to variations in class hardness. One way to quantify class hardness is through sample complexity - the minimum number of samples required to ef...
2504.07061
Shi Pan
Shi Pan and Jianan Chen and Maria Secrier
Teaching pathology foundation models to accurately predict gene expression with parameter efficient knowledge transfer
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Gene expression profiling provides critical insights into cellular heterogeneity, biological processes and disease mechanisms. There has been an increasing interest in computational approaches that can predict gene expression directly from digitalized histopathology images. While image foundation models have shown pr...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 17:24:41 GMT" } ]
2025-04-10T00:00:00
[ [ "Pan", "Shi", "" ], [ "Chen", "Jianan", "" ], [ "Secrier", "Maria", "" ] ]
TITLE: Teaching pathology foundation models to accurately predict gene expression with parameter efficient knowledge transfer ABSTRACT: Gene expression profiling provides critical insights into cellular heterogeneity, biological processes and disease mechanisms. There has been an increasing interest in computatio...
2504.07065
William Simon
Riselda Kodra, Hadjer Benmeziane, Irem Boybat, William Andrew Simon
Enhancing Downstream Analysis in Genome Sequencing: Species Classification While Basecalling
Accepted as Tiny Paper at MLGenX workshop, ICLR, 2025
null
null
null
q-bio.GN cs.LG
http://creativecommons.org/licenses/by/4.0/
The ability to quickly and accurately identify microbial species in a sample, known as metagenomic profiling, is critical across various fields, from healthcare to environmental science. This paper introduces a novel method to profile signals coming from sequencing devices in parallel with determining their nucleotid...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 17:30:43 GMT" } ]
2025-04-10T00:00:00
[ [ "Kodra", "Riselda", "" ], [ "Benmeziane", "Hadjer", "" ], [ "Boybat", "Irem", "" ], [ "Simon", "William Andrew", "" ] ]
TITLE: Enhancing Downstream Analysis in Genome Sequencing: Species Classification While Basecalling ABSTRACT: The ability to quickly and accurately identify microbial species in a sample, known as metagenomic profiling, is critical across various fields, from healthcare to environmental science. This paper introd...
2504.07069
Bibek Paudel
Bibek Paudel, Alexander Lyzhov, Preetam Joshi, Puneet Anand
HalluciNot: Hallucination Detection Through Context and Common Knowledge Verification
null
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by-sa/4.0/
This paper introduces a comprehensive system for detecting hallucinations in large language model (LLM) outputs in enterprise settings. We present a novel taxonomy of LLM responses specific to hallucination in enterprise applications, categorizing them into context-based, common knowledge, enterprise-specific, and in...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 17:39:41 GMT" } ]
2025-04-10T00:00:00
[ [ "Paudel", "Bibek", "" ], [ "Lyzhov", "Alexander", "" ], [ "Joshi", "Preetam", "" ], [ "Anand", "Puneet", "" ] ]
TITLE: HalluciNot: Hallucination Detection Through Context and Common Knowledge Verification ABSTRACT: This paper introduces a comprehensive system for detecting hallucinations in large language model (LLM) outputs in enterprise settings. We present a novel taxonomy of LLM responses specific to hallucination in e...
2504.07072
Desmond Elliott
Israfel Salazar, Manuel Fern\'andez Burda, Shayekh Bin Islam, Arshia Soltani Moakhar, Shivalika Singh, Fabian Farestam, Angelika Romanou, Danylo Boiko, Dipika Khullar, Mike Zhang, Dominik Krzemi\'nski, Jekaterina Novikova, Lu\'isa Shimabucoro, Joseph Marvin Imperial, Rishabh Maheshwary, Sharad Duwal, Alfonso Am...
Kaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation
null
null
null
null
cs.CL cs.CV
http://creativecommons.org/licenses/by/4.0/
The evaluation of vision-language models (VLMs) has mainly relied on English-language benchmarks, leaving significant gaps in both multilingual and multicultural coverage. While multilingual benchmarks have expanded, both in size and languages, many rely on translations of English datasets, failing to capture cultura...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 17:43:16 GMT" } ]
2025-04-10T00:00:00
[ [ "Salazar", "Israfel", "" ], [ "Burda", "Manuel Fernández", "" ], [ "Islam", "Shayekh Bin", "" ], [ "Moakhar", "Arshia Soltani", "" ], [ "Singh", "Shivalika", "" ], [ "Farestam", "Fabian", "" ], [ "Romanou", "An...
TITLE: Kaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation ABSTRACT: The evaluation of vision-language models (VLMs) has mainly relied on English-language benchmarks, leaving significant gaps in both multilingual and multicultural coverage. While multilingual benchmarks have expanded, bot...
2504.07080
Atharva Pandey
Atharva Pandey, Kshitij Dubey, Rahul Sharma, Amit Sharma
DeduCE: Deductive Consistency as a Framework to Evaluate LLM Reasoning
null
null
null
null
cs.CL cs.AI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Despite great performance on Olympiad-level reasoning problems, frontier large language models can still struggle on high school math when presented with novel problems outside standard benchmarks. Going beyond final accuracy, we propose a deductive consistency metric to analyze chain-of-thought output from language ...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 17:53:55 GMT" } ]
2025-04-10T00:00:00
[ [ "Pandey", "Atharva", "" ], [ "Dubey", "Kshitij", "" ], [ "Sharma", "Rahul", "" ], [ "Sharma", "Amit", "" ] ]
TITLE: DeduCE: Deductive Consistency as a Framework to Evaluate LLM Reasoning ABSTRACT: Despite great performance on Olympiad-level reasoning problems, frontier large language models can still struggle on high school math when presented with novel problems outside standard benchmarks. Going beyond final accuracy, w...
2504.07093
Gene Chou
Gene Chou, Wenqi Xian, Guandao Yang, Mohamed Abdelfattah, Bharath Hariharan, Noah Snavely, Ning Yu, Paul Debevec
FlashDepth: Real-time Streaming Video Depth Estimation at 2K Resolution
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A versatile video depth estimation model should (1) be accurate and consistent across frames, (2) produce high-resolution depth maps, and (3) support real-time streaming. We propose FlashDepth, a method that satisfies all three requirements, performing depth estimation on a 2044x1148 streaming video at 24 FPS. We sho...
[ { "version": "v1", "created": "Wed, 9 Apr 2025 17:59:31 GMT" } ]
2025-04-10T00:00:00
[ [ "Chou", "Gene", "" ], [ "Xian", "Wenqi", "" ], [ "Yang", "Guandao", "" ], [ "Abdelfattah", "Mohamed", "" ], [ "Hariharan", "Bharath", "" ], [ "Snavely", "Noah", "" ], [ "Yu", "Ning", "" ], [ "Debeve...
TITLE: FlashDepth: Real-time Streaming Video Depth Estimation at 2K Resolution ABSTRACT: A versatile video depth estimation model should (1) be accurate and consistent across frames, (2) produce high-resolution depth maps, and (3) support real-time streaming. We propose FlashDepth, a method that satisfies all three...
2110.03427
Atanu Mandal
Atanu Mandal, Santanu Pal, Indranil Dutta, Mahidas Bhattacharya, Sudip Kumar Naskar
Is Attention always needed? A Case Study on Language Identification from Speech
Accepted for publication in Natural Language Engineering
Nat. lang. process. 31 (2025) 250-276
10.1017/nlp.2024.22
null
cs.LG cs.CL cs.SD eess.AS eess.SP
http://creativecommons.org/licenses/by/4.0/
Language Identification (LID) is a crucial preliminary process in the field of Automatic Speech Recognition (ASR) that involves the identification of a spoken language from audio samples. Contemporary systems that can process speech in multiple languages require users to expressly designate one or more languages prio...
[ { "version": "v1", "created": "Tue, 5 Oct 2021 16:38:57 GMT" }, { "version": "v2", "created": "Sun, 10 Jul 2022 03:47:05 GMT" }, { "version": "v3", "created": "Wed, 25 Oct 2023 15:21:08 GMT" } ]
2025-04-09T00:00:00
[ [ "Mandal", "Atanu", "" ], [ "Pal", "Santanu", "" ], [ "Dutta", "Indranil", "" ], [ "Bhattacharya", "Mahidas", "" ], [ "Naskar", "Sudip Kumar", "" ] ]
TITLE: Is Attention always needed? A Case Study on Language Identification from Speech ABSTRACT: Language Identification (LID) is a crucial preliminary process in the field of Automatic Speech Recognition (ASR) that involves the identification of a spoken language from audio samples. Contemporary systems that can...
2111.13463
Ivica Kostric
Ivica Kostric and Krisztian Balog and Filip Radlinski
Generating Usage-related Questions for Preference Elicitation in Conversational Recommender Systems
Journal extension of our RecSys '21 paper titled "Soliciting User Preferences in Conversational Recommender Systems via Usage-related Questions." This version appears in ACM Transactions on Recommender Systems (ToRS), 2(2), Article 12, April 2024, with expanded experiments and new analysis
ACM Transactions on Recommender Systems (ToRS), Volume 2, Issue 2, Article 12 (April 2024)
10.1145/3629981
null
cs.IR cs.AI cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
A key distinguishing feature of conversational recommender systems over traditional recommender systems is their ability to elicit user preferences using natural language. Currently, the predominant approach to preference elicitation is to ask questions directly about items or item attributes. Users searching for rec...
[ { "version": "v1", "created": "Fri, 26 Nov 2021 12:23:14 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 13:25:51 GMT" } ]
2025-04-09T00:00:00
[ [ "Kostric", "Ivica", "" ], [ "Balog", "Krisztian", "" ], [ "Radlinski", "Filip", "" ] ]
TITLE: Generating Usage-related Questions for Preference Elicitation in Conversational Recommender Systems ABSTRACT: A key distinguishing feature of conversational recommender systems over traditional recommender systems is their ability to elicit user preferences using natural language. Currently, the predominan...
2210.15527
Yun-Hin Chan
Yun-Hin Chan, Edith C.-H. Ngai
Exploiting Features and Logits in Heterogeneous Federated Learning
Accepted by Computer Networks
null
10.1016/j.comnet.2025.111271
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Due to the rapid growth of IoT and artificial intelligence, deploying neural networks on IoT devices is becoming increasingly crucial for edge intelligence. Federated learning (FL) facilitates the management of edge devices to collaboratively train a shared model while maintaining training data local and private. How...
[ { "version": "v1", "created": "Thu, 27 Oct 2022 15:11:46 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 09:54:58 GMT" } ]
2025-04-09T00:00:00
[ [ "Chan", "Yun-Hin", "" ], [ "Ngai", "Edith C. -H.", "" ] ]
TITLE: Exploiting Features and Logits in Heterogeneous Federated Learning ABSTRACT: Due to the rapid growth of IoT and artificial intelligence, deploying neural networks on IoT devices is becoming increasingly crucial for edge intelligence. Federated learning (FL) facilitates the management of edge devices to colla...
2301.00539
Sudhansu Bala Das
Sudhansu Bala Das, Divyajoti Panda, Tapas Kumar Mishra, Bidyut Kr. Patra
Statistical Machine Translation for Indic Languages
32pages, 1 figure, 4 tables
Nat. lang. process. 31 (2025) 328-345
10.1017/nlp.2024.26
null
cs.CL
http://creativecommons.org/licenses/by/4.0/
Machine Translation (MT) system generally aims at automatic representation of source language into target language retaining the originality of context using various Natural Language Processing (NLP) techniques. Among various NLP methods, Statistical Machine Translation(SMT). SMT uses probabilistic and statistical te...
[ { "version": "v1", "created": "Mon, 2 Jan 2023 06:23:12 GMT" } ]
2025-04-09T00:00:00
[ [ "Das", "Sudhansu Bala", "" ], [ "Panda", "Divyajoti", "" ], [ "Mishra", "Tapas Kumar", "" ], [ "Patra", "Bidyut Kr.", "" ] ]
TITLE: Statistical Machine Translation for Indic Languages ABSTRACT: Machine Translation (MT) system generally aims at automatic representation of source language into target language retaining the originality of context using various Natural Language Processing (NLP) techniques. Among various NLP methods, Statisti...
2301.06650
Lijun Sun Dr.
Vincent Zhihao Zheng, Seongjin Choi, Lijun Sun
Probabilistic Traffic Forecasting with Dynamic Regression
null
Probabilistic Traffic Forecasting with Dynamic Regression. Transportation Science (2025)
10.1287/trsc.2024.0560
null
cs.LG stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper proposes a dynamic regression (DR) framework that enhances existing deep spatiotemporal models by incorporating structured learning for the error process in traffic forecasting. The framework relaxes the assumption of time independence by modeling the error series of the base model (i.e., a well-establishe...
[ { "version": "v1", "created": "Tue, 17 Jan 2023 01:12:44 GMT" }, { "version": "v2", "created": "Fri, 31 May 2024 15:05:40 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 14:26:10 GMT" } ]
2025-04-09T00:00:00
[ [ "Zheng", "Vincent Zhihao", "" ], [ "Choi", "Seongjin", "" ], [ "Sun", "Lijun", "" ] ]
TITLE: Probabilistic Traffic Forecasting with Dynamic Regression ABSTRACT: This paper proposes a dynamic regression (DR) framework that enhances existing deep spatiotemporal models by incorporating structured learning for the error process in traffic forecasting. The framework relaxes the assumption of time indepen...
2305.15203
Lorenzo Basile
Lorenzo Basile, Nikos Karantzas, Alberto d'Onofrio, Luca Manzoni, Luca Bortolussi, Alex Rodriguez, Fabio Anselmi
Frequency maps reveal the correlation between Adversarial Attacks and Implicit Bias
Accepted at IJCNN 2025
null
null
null
cs.LG cs.AI cs.CR stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Despite their impressive performance in classification tasks, neural networks are known to be vulnerable to adversarial attacks, subtle perturbations of the input data designed to deceive the model. In this work, we investigate the correlation between these perturbations and the implicit bias of neural networks train...
[ { "version": "v1", "created": "Wed, 24 May 2023 14:40:23 GMT" }, { "version": "v2", "created": "Wed, 17 Jul 2024 16:34:48 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 14:29:39 GMT" } ]
2025-04-09T00:00:00
[ [ "Basile", "Lorenzo", "" ], [ "Karantzas", "Nikos", "" ], [ "d'Onofrio", "Alberto", "" ], [ "Manzoni", "Luca", "" ], [ "Bortolussi", "Luca", "" ], [ "Rodriguez", "Alex", "" ], [ "Anselmi", "Fabio", "" ] ]
TITLE: Frequency maps reveal the correlation between Adversarial Attacks and Implicit Bias ABSTRACT: Despite their impressive performance in classification tasks, neural networks are known to be vulnerable to adversarial attacks, subtle perturbations of the input data designed to deceive the model. In this work, ...
2310.16810
Yongxin Zhou
Yongxin Zhou, Fabien Ringeval, Fran\c{c}ois Portet
Can GPT models Follow Human Summarization Guidelines? A Study for Targeted Communication Goals
null
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by-nc-sa/4.0/
This study investigates the ability of GPT models (ChatGPT, GPT-4 and GPT-4o) to generate dialogue summaries that adhere to human guidelines. Our evaluation involved experimenting with various prompts to guide the models in complying with guidelines on two datasets: DialogSum (English social conversations) and DECODA...
[ { "version": "v1", "created": "Wed, 25 Oct 2023 17:39:07 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 21:42:15 GMT" } ]
2025-04-09T00:00:00
[ [ "Zhou", "Yongxin", "" ], [ "Ringeval", "Fabien", "" ], [ "Portet", "François", "" ] ]
TITLE: Can GPT models Follow Human Summarization Guidelines? A Study for Targeted Communication Goals ABSTRACT: This study investigates the ability of GPT models (ChatGPT, GPT-4 and GPT-4o) to generate dialogue summaries that adhere to human guidelines. Our evaluation involved experimenting with various prompts t...
2311.01759
Jianlei Yang
Jianlei Yang, Jiacheng Liao, Fanding Lei, Meichen Liu, Junyi Chen, Lingkun Long, Han Wan, Bei Yu, Weisheng Zhao
TinyFormer: Efficient Transformer Design and Deployment on Tiny Devices
This work has been submitted to the IEEE for possible publication
null
null
null
cs.LG cs.AR
http://creativecommons.org/licenses/by-nc-nd/4.0/
Developing deep learning models on tiny devices (e.g. Microcontroller units, MCUs) has attracted much attention in various embedded IoT applications. However, it is challenging to efficiently design and deploy recent advanced models (e.g. transformers) on tiny devices due to their severe hardware resource constraints...
[ { "version": "v1", "created": "Fri, 3 Nov 2023 07:34:47 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 11:42:15 GMT" } ]
2025-04-09T00:00:00
[ [ "Yang", "Jianlei", "" ], [ "Liao", "Jiacheng", "" ], [ "Lei", "Fanding", "" ], [ "Liu", "Meichen", "" ], [ "Chen", "Junyi", "" ], [ "Long", "Lingkun", "" ], [ "Wan", "Han", "" ], [ "Yu", "Bei", ...
TITLE: TinyFormer: Efficient Transformer Design and Deployment on Tiny Devices ABSTRACT: Developing deep learning models on tiny devices (e.g. Microcontroller units, MCUs) has attracted much attention in various embedded IoT applications. However, it is challenging to efficiently design and deploy recent advanced m...
2311.18681
Chantal Pellegrini
Chantal Pellegrini, Ege \"Ozsoy, Benjamin Busam, Nassir Navab, Matthias Keicher
RaDialog: A Large Vision-Language Model for Radiology Report Generation and Conversational Assistance
improved version accepted at MIDL 2025: https://openreview.net/pdf?id=trUvr1gSNI
null
null
null
cs.CV cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Conversational AI tools that can generate and discuss clinically correct radiology reports for a given medical image have the potential to transform radiology. Such a human-in-the-loop radiology assistant could facilitate a collaborative diagnostic process, thus saving time and improving the quality of reports. Towar...
[ { "version": "v1", "created": "Thu, 30 Nov 2023 16:28:40 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 07:32:34 GMT" } ]
2025-04-09T00:00:00
[ [ "Pellegrini", "Chantal", "" ], [ "Özsoy", "Ege", "" ], [ "Busam", "Benjamin", "" ], [ "Navab", "Nassir", "" ], [ "Keicher", "Matthias", "" ] ]
TITLE: RaDialog: A Large Vision-Language Model for Radiology Report Generation and Conversational Assistance ABSTRACT: Conversational AI tools that can generate and discuss clinically correct radiology reports for a given medical image have the potential to transform radiology. Such a human-in-the-loop radiology ...
2312.16379
Alexey Melnikov
Asel Sagingalieva, Stefan Komornyik, Ayush Joshi, Christopher Mansell, Karan Pinto, Markus Pflitsch, and Alexey Melnikov
Photovoltaic power forecasting using quantum machine learning
12 pages, 4 figures, 1 table
null
null
null
cs.LG cs.ET quant-ph
http://creativecommons.org/licenses/by/4.0/
Predicting solar panel power output is crucial for advancing the transition to renewable energy but is complicated by the variable and non-linear nature of solar energy. This is influenced by numerous meteorological factors, geographical positioning, and photovoltaic cell properties, posing significant challenges to ...
[ { "version": "v1", "created": "Wed, 27 Dec 2023 02:37:46 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 22:55:21 GMT" } ]
2025-04-09T00:00:00
[ [ "Sagingalieva", "Asel", "" ], [ "Komornyik", "Stefan", "" ], [ "Joshi", "Ayush", "" ], [ "Mansell", "Christopher", "" ], [ "Pinto", "Karan", "" ], [ "Pflitsch", "Markus", "" ], [ "Melnikov", "Alexey", "" ...
TITLE: Photovoltaic power forecasting using quantum machine learning ABSTRACT: Predicting solar panel power output is crucial for advancing the transition to renewable energy but is complicated by the variable and non-linear nature of solar energy. This is influenced by numerous meteorological factors, geographical...
2402.04051
Akira Ito
Akira Ito, Masanori Yamada, Atsutoshi Kumagai
Analysis of Linear Mode Connectivity via Permutation-Based Weight Matching: With Insights into Other Permutation Search Methods
In Proceedings of the Thirteenth International Conference on Learning Representations (ICLR 2025)
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recently, Ainsworth et al. showed that using weight matching (WM) to minimize the $L^2$ distance in a permutation search of model parameters effectively identifies permutations that satisfy linear mode connectivity (LMC), where the loss along a linear path between two independently trained models with different seeds...
[ { "version": "v1", "created": "Tue, 6 Feb 2024 14:53:28 GMT" }, { "version": "v2", "created": "Mon, 19 Feb 2024 10:36:25 GMT" }, { "version": "v3", "created": "Mon, 15 Apr 2024 05:57:26 GMT" }, { "version": "v4", "created": "Thu, 3 Oct 2024 11:36:28 GMT" }, { "ver...
2025-04-09T00:00:00
[ [ "Ito", "Akira", "" ], [ "Yamada", "Masanori", "" ], [ "Kumagai", "Atsutoshi", "" ] ]
TITLE: Analysis of Linear Mode Connectivity via Permutation-Based Weight Matching: With Insights into Other Permutation Search Methods ABSTRACT: Recently, Ainsworth et al. showed that using weight matching (WM) to minimize the $L^2$ distance in a permutation search of model parameters effectively identifies permu...
2403.02437
Hyejun Jeong
Hyejun Jeong, Shiqing Ma, Amir Houmansadr
A Survey on Federated Unlearning: Challenges and Opportunities
null
null
null
null
cs.LG cs.AI cs.DC
http://creativecommons.org/licenses/by/4.0/
Federated learning (FL), introduced in 2017, facilitates collaborative learning between non-trusting parties with no need for the parties to explicitly share their data among themselves. This allows training models on user data while respecting privacy regulations such as GDPR and CPRA. However, emerging privacy requ...
[ { "version": "v1", "created": "Mon, 4 Mar 2024 19:35:08 GMT" }, { "version": "v2", "created": "Wed, 5 Jun 2024 19:00:03 GMT" }, { "version": "v3", "created": "Mon, 7 Apr 2025 19:55:57 GMT" } ]
2025-04-09T00:00:00
[ [ "Jeong", "Hyejun", "" ], [ "Ma", "Shiqing", "" ], [ "Houmansadr", "Amir", "" ] ]
TITLE: A Survey on Federated Unlearning: Challenges and Opportunities ABSTRACT: Federated learning (FL), introduced in 2017, facilitates collaborative learning between non-trusting parties with no need for the parties to explicitly share their data among themselves. This allows training models on user data while re...
2404.03543
JiaWei Guo
Jiawei Guo, Ziming Li, Xueling Liu, Kaijing Ma, Tianyu Zheng, Zhouliang Yu, Ding Pan, Yizhi LI, Ruibo Liu, Yue Wang, Shuyue Guo, Xingwei Qu, Xiang Yue, Ge Zhang, Wenhu Chen, Jie Fu
CodeEditorBench: Evaluating Code Editing Capability of Large Language Models
null
null
null
null
cs.SE cs.AI cs.CL cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Large Language Models (LLMs) for code are rapidly evolving, with code editing emerging as a critical capability. We introduce CodeEditorBench, an evaluation framework designed to rigorously assess the performance of LLMs in code editing tasks, including debugging, translating, polishing, and requirement switching. Un...
[ { "version": "v1", "created": "Thu, 4 Apr 2024 15:49:49 GMT" }, { "version": "v2", "created": "Sat, 6 Apr 2024 04:29:25 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 09:39:25 GMT" } ]
2025-04-09T00:00:00
[ [ "Guo", "Jiawei", "" ], [ "Li", "Ziming", "" ], [ "Liu", "Xueling", "" ], [ "Ma", "Kaijing", "" ], [ "Zheng", "Tianyu", "" ], [ "Yu", "Zhouliang", "" ], [ "Pan", "Ding", "" ], [ "LI", "Yizhi", ...
TITLE: CodeEditorBench: Evaluating Code Editing Capability of Large Language Models ABSTRACT: Large Language Models (LLMs) for code are rapidly evolving, with code editing emerging as a critical capability. We introduce CodeEditorBench, an evaluation framework designed to rigorously assess the performance of LLMs...
2405.10577
Yizhe Zhao
Zhe Huang, Yizhe Zhao, Hao Xiao, Chenyan Wu, Lingting Ge
DuoSpaceNet: Leveraging Both Bird's-Eye-View and Perspective View Representations for 3D Object Detection
CVPR 2025 Workshop on Autonomous Driving (WAD)
null
null
null
cs.CV cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Multi-view camera-only 3D object detection largely follows two primary paradigms: exploiting bird's-eye-view (BEV) representations or focusing on perspective-view (PV) features, each with distinct advantages. Although several recent approaches explore combining BEV and PV, many rely on partial fusion or maintain sepa...
[ { "version": "v1", "created": "Fri, 17 May 2024 07:04:29 GMT" }, { "version": "v2", "created": "Thu, 29 Aug 2024 02:09:11 GMT" }, { "version": "v3", "created": "Mon, 7 Apr 2025 18:00:17 GMT" } ]
2025-04-09T00:00:00
[ [ "Huang", "Zhe", "" ], [ "Zhao", "Yizhe", "" ], [ "Xiao", "Hao", "" ], [ "Wu", "Chenyan", "" ], [ "Ge", "Lingting", "" ] ]
TITLE: DuoSpaceNet: Leveraging Both Bird's-Eye-View and Perspective View Representations for 3D Object Detection ABSTRACT: Multi-view camera-only 3D object detection largely follows two primary paradigms: exploiting bird's-eye-view (BEV) representations or focusing on perspective-view (PV) features, each with dis...
2405.13955
Xiaoshan Zhou
Xiaoshan Zhou, Carol C. Menassa, and Vineet R. Kamat
Decoding Brain Dynamics in Motor Planning Based on EEG Microstates for Predicting Pedestrian Road-Crossing in Vehicle-to-Everything Architectures
38 pages, 11 figures
null
null
null
cs.HC cs.ET
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Pedestrians who cross roads, often emerge from occlusion or abruptly begin crossing from a standstill, frequently leading to unintended collisions with vehicular traffic that result in accidents and interruptions. Existing studies have predominantly relied on external network sensing and observational data to anticip...
[ { "version": "v1", "created": "Wed, 22 May 2024 19:40:37 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 19:58:30 GMT" } ]
2025-04-09T00:00:00
[ [ "Zhou", "Xiaoshan", "" ], [ "Menassa", "Carol C.", "" ], [ "Kamat", "Vineet R.", "" ] ]
TITLE: Decoding Brain Dynamics in Motor Planning Based on EEG Microstates for Predicting Pedestrian Road-Crossing in Vehicle-to-Everything Architectures ABSTRACT: Pedestrians who cross roads, often emerge from occlusion or abruptly begin crossing from a standstill, frequently leading to unintended collisions with...
2405.13983
Anton Morgunov
Yu Shee, Anton Morgunov, Haote Li, Victor S. Batista
DirectMultiStep: Direct Route Generation for Multistep Retrosynthesis
null
null
10.1021/acs.jcim.4c01982
null
cs.LG cs.AI
http://creativecommons.org/licenses/by/4.0/
Traditional computer-aided synthesis planning (CASP) methods rely on iterative single-step predictions, leading to exponential search space growth that limits efficiency and scalability. We introduce a series of transformer-based models, that leverage a mixture of experts approach to directly generate multistep synth...
[ { "version": "v1", "created": "Wed, 22 May 2024 20:39:05 GMT" }, { "version": "v2", "created": "Tue, 21 Jan 2025 17:37:07 GMT" }, { "version": "v3", "created": "Thu, 20 Mar 2025 01:58:12 GMT" } ]
2025-04-09T00:00:00
[ [ "Shee", "Yu", "" ], [ "Morgunov", "Anton", "" ], [ "Li", "Haote", "" ], [ "Batista", "Victor S.", "" ] ]
TITLE: DirectMultiStep: Direct Route Generation for Multistep Retrosynthesis ABSTRACT: Traditional computer-aided synthesis planning (CASP) methods rely on iterative single-step predictions, leading to exponential search space growth that limits efficiency and scalability. We introduce a series of transformer-based...
2405.20445
Jianan Zhao
Jianan Zhao, Zhaocheng Zhu, Mikhail Galkin, Hesham Mostafa, Michael Bronstein, Jian Tang
Fully-inductive Node Classification on Arbitrary Graphs
ICLR2025
null
null
null
cs.LG cs.SI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
One fundamental challenge in graph machine learning is generalizing to new graphs. Many existing methods following the inductive setup can generalize to test graphs with new structures, but assuming the feature and label spaces remain the same as the training ones. This paper introduces a fully-inductive setup, where...
[ { "version": "v1", "created": "Thu, 30 May 2024 19:43:29 GMT" }, { "version": "v2", "created": "Mon, 3 Jun 2024 02:08:54 GMT" }, { "version": "v3", "created": "Sun, 9 Feb 2025 03:14:20 GMT" }, { "version": "v4", "created": "Fri, 28 Feb 2025 00:56:45 GMT" }, { "ver...
2025-04-09T00:00:00
[ [ "Zhao", "Jianan", "" ], [ "Zhu", "Zhaocheng", "" ], [ "Galkin", "Mikhail", "" ], [ "Mostafa", "Hesham", "" ], [ "Bronstein", "Michael", "" ], [ "Tang", "Jian", "" ] ]
TITLE: Fully-inductive Node Classification on Arbitrary Graphs ABSTRACT: One fundamental challenge in graph machine learning is generalizing to new graphs. Many existing methods following the inductive setup can generalize to test graphs with new structures, but assuming the feature and label spaces remain the same...
2405.20769
Matthew Regehr
Christian Janos Lebeda, Matthew Regehr, Gautam Kamath, Thomas Steinke
Avoiding Pitfalls for Privacy Accounting of Subsampled Mechanisms under Composition
null
null
null
null
cs.CR cs.DS cs.LG stat.ML
http://creativecommons.org/licenses/by/4.0/
We consider the problem of computing tight privacy guarantees for the composition of subsampled differentially private mechanisms. Recent algorithms can numerically compute the privacy parameters to arbitrary precision but must be carefully applied. Our main contribution is to address two common points of confusion...
[ { "version": "v1", "created": "Mon, 27 May 2024 20:30:12 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 15:21:03 GMT" } ]
2025-04-09T00:00:00
[ [ "Lebeda", "Christian Janos", "" ], [ "Regehr", "Matthew", "" ], [ "Kamath", "Gautam", "" ], [ "Steinke", "Thomas", "" ] ]
TITLE: Avoiding Pitfalls for Privacy Accounting of Subsampled Mechanisms under Composition ABSTRACT: We consider the problem of computing tight privacy guarantees for the composition of subsampled differentially private mechanisms. Recent algorithms can numerically compute the privacy parameters to arbitrary prec...
2406.00984
Hiroaki Yamagiwa
Hiroaki Yamagiwa, Ryoma Hashimoto, Kiwamu Arakane, Ken Murakami, Shou Soeda, Momose Oyama, Yihua Zhu, Mariko Okada, Hidetoshi Shimodaira
Predicting Drug-Gene Relations via Analogy Tasks with Word Embeddings
null
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Natural language processing (NLP) is utilized in a wide range of fields, where words in text are typically transformed into feature vectors called embeddings. BioConceptVec is a specific example of embeddings tailored for biology, trained on approximately 30 million PubMed abstracts using models such as skip-gram. Ge...
[ { "version": "v1", "created": "Mon, 3 Jun 2024 04:36:38 GMT" }, { "version": "v2", "created": "Wed, 4 Sep 2024 20:22:41 GMT" }, { "version": "v3", "created": "Sun, 8 Dec 2024 09:03:03 GMT" }, { "version": "v4", "created": "Tue, 8 Apr 2025 17:50:27 GMT" } ]
2025-04-09T00:00:00
[ [ "Yamagiwa", "Hiroaki", "" ], [ "Hashimoto", "Ryoma", "" ], [ "Arakane", "Kiwamu", "" ], [ "Murakami", "Ken", "" ], [ "Soeda", "Shou", "" ], [ "Oyama", "Momose", "" ], [ "Zhu", "Yihua", "" ], [ "Okad...
TITLE: Predicting Drug-Gene Relations via Analogy Tasks with Word Embeddings ABSTRACT: Natural language processing (NLP) is utilized in a wide range of fields, where words in text are typically transformed into feature vectors called embeddings. BioConceptVec is a specific example of embeddings tailored for biology...
2406.07467
Fatemeh Hadadi
Fatemeh Hadadi, Qinghua Xu, Domenico Bianculli, Lionel Briand
LLM meets ML: Data-efficient Anomaly Detection on Unseen Unstable Logs
null
null
null
null
cs.SE
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Most log-based anomaly detectors assume logs are stable, though logs are often unstable due to software or environmental changes. Anomaly detection on unstable logs (ULAD) is therefore a more realistic, yet under-investigated challenge. Current approaches predominantly employ machine learning (ML) models, which often...
[ { "version": "v1", "created": "Tue, 11 Jun 2024 17:13:18 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 20:52:04 GMT" } ]
2025-04-09T00:00:00
[ [ "Hadadi", "Fatemeh", "" ], [ "Xu", "Qinghua", "" ], [ "Bianculli", "Domenico", "" ], [ "Briand", "Lionel", "" ] ]
TITLE: LLM meets ML: Data-efficient Anomaly Detection on Unseen Unstable Logs ABSTRACT: Most log-based anomaly detectors assume logs are stable, though logs are often unstable due to software or environmental changes. Anomaly detection on unstable logs (ULAD) is therefore a more realistic, yet under-investigated ch...
2406.08092
Zhi Qu
Zhi Qu, Chenchen Ding, Taro Watanabe
Languages Transferred Within the Encoder: On Representation Transfer in Zero-Shot Multilingual Translation
Accepted by MT Summit 2025
null
null
null
cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Understanding representation transfer in multilingual neural machine translation (MNMT) can reveal the reason for the zero-shot translation deficiency. In this work, we systematically analyze the representational issue of MNMT models. We first introduce the identity pair, translating a sentence to itself, to address ...
[ { "version": "v1", "created": "Wed, 12 Jun 2024 11:16:30 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 03:39:51 GMT" } ]
2025-04-09T00:00:00
[ [ "Qu", "Zhi", "" ], [ "Ding", "Chenchen", "" ], [ "Watanabe", "Taro", "" ] ]
TITLE: Languages Transferred Within the Encoder: On Representation Transfer in Zero-Shot Multilingual Translation ABSTRACT: Understanding representation transfer in multilingual neural machine translation (MNMT) can reveal the reason for the zero-shot translation deficiency. In this work, we systematically analyz...
2406.11917
Chao He
Chao He and Hongmei Shi and Ruixin Li and Jianbo Li and ZuJun Yu
Modulated Differentiable STFT and Balanced Spectrum Metric for Freight Train Wheelset Bearing Cross-machine Transfer Fault Diagnosis under Speed Fluctuations
null
Advanced Engineering Informatics 62 (2024) 102568
10.1016/j.aei.2024.102568
null
cs.LG eess.SP
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The service conditions of wheelset bearings has a direct impact on the safe operation of railway heavy haul freight trains as the key components. However, speed fluctuation of the trains and few fault samples are the two main problems that restrict the accuracy of bearing fault diagnosis. Therefore, a cross-machine t...
[ { "version": "v1", "created": "Mon, 17 Jun 2024 02:43:24 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 04:01:43 GMT" } ]
2025-04-09T00:00:00
[ [ "He", "Chao", "" ], [ "Shi", "Hongmei", "" ], [ "Li", "Ruixin", "" ], [ "Li", "Jianbo", "" ], [ "Yu", "ZuJun", "" ] ]
TITLE: Modulated Differentiable STFT and Balanced Spectrum Metric for Freight Train Wheelset Bearing Cross-machine Transfer Fault Diagnosis under Speed Fluctuations ABSTRACT: The service conditions of wheelset bearings has a direct impact on the safe operation of railway heavy haul freight trains as the key com...
2406.15341
Haoyang Liu
Haoyang Liu, Shuyu Chen, Ye Zhang, Haohan Wang
GenoTEX: An LLM Agent Benchmark for Automated Gene Expression Data Analysis
31 pages, 4 figures
null
null
null
cs.LG cs.AI q-bio.GN
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recent advancements in machine learning have significantly improved the identification of disease-associated genes from gene expression datasets. However, these processes often require extensive expertise and manual effort, limiting their scalability. Large Language Model (LLM)-based agents have shown promise in auto...
[ { "version": "v1", "created": "Fri, 21 Jun 2024 17:55:24 GMT" }, { "version": "v2", "created": "Thu, 27 Mar 2025 17:59:22 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 17:09:04 GMT" } ]
2025-04-09T00:00:00
[ [ "Liu", "Haoyang", "" ], [ "Chen", "Shuyu", "" ], [ "Zhang", "Ye", "" ], [ "Wang", "Haohan", "" ] ]
TITLE: GenoTEX: An LLM Agent Benchmark for Automated Gene Expression Data Analysis ABSTRACT: Recent advancements in machine learning have significantly improved the identification of disease-associated genes from gene expression datasets. However, these processes often require extensive expertise and manual effor...
2407.21077
Vahid Noroozi
Somshubra Majumdar, Vahid Noroozi, Mehrzad Samadi, Sean Narenthiran, Aleksander Ficek, Wasi Uddin Ahmad, Jocelyn Huang, Jagadeesh Balam, Boris Ginsburg
Genetic Instruct: Scaling up Synthetic Generation of Coding Instructions for Large Language Models
null
null
null
null
cs.CL cs.LG cs.NE
http://creativecommons.org/licenses/by/4.0/
Large Language Models (LLMs) require high quality instruction data for effective alignment, particularly in code generation tasks where expert curated datasets are expensive to produce. We present Genetic-Instruct, a scalable algorithm for synthesizing large-scale, high quality coding instructions using evolutionary ...
[ { "version": "v1", "created": "Mon, 29 Jul 2024 20:42:59 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 23:35:11 GMT" } ]
2025-04-09T00:00:00
[ [ "Majumdar", "Somshubra", "" ], [ "Noroozi", "Vahid", "" ], [ "Samadi", "Mehrzad", "" ], [ "Narenthiran", "Sean", "" ], [ "Ficek", "Aleksander", "" ], [ "Ahmad", "Wasi Uddin", "" ], [ "Huang", "Jocelyn", "" ...
TITLE: Genetic Instruct: Scaling up Synthetic Generation of Coding Instructions for Large Language Models ABSTRACT: Large Language Models (LLMs) require high quality instruction data for effective alignment, particularly in code generation tasks where expert curated datasets are expensive to produce. We present G...
2408.04290
Amirreza Fateh
Alireza Saber, Pouria Parhami, Alimohammad Siahkarzadeh, Mansoor Fateh, Amirreza Fateh
Efficient and Accurate Pneumonia Detection Using a Novel Multi-Scale Transformer Approach
null
null
null
null
eess.IV cs.CV cs.LG
http://creativecommons.org/licenses/by/4.0/
Pneumonia, a prevalent respiratory infection, remains a leading cause of morbidity and mortality worldwide, particularly among vulnerable populations. Chest X-rays serve as a primary tool for pneumonia detection; however, variations in imaging conditions and subtle visual indicators complicate consistent interpretati...
[ { "version": "v1", "created": "Thu, 8 Aug 2024 08:06:42 GMT" }, { "version": "v2", "created": "Sun, 3 Nov 2024 11:51:50 GMT" }, { "version": "v3", "created": "Sun, 26 Jan 2025 17:04:30 GMT" }, { "version": "v4", "created": "Tue, 8 Apr 2025 07:00:02 GMT" } ]
2025-04-09T00:00:00
[ [ "Saber", "Alireza", "" ], [ "Parhami", "Pouria", "" ], [ "Siahkarzadeh", "Alimohammad", "" ], [ "Fateh", "Mansoor", "" ], [ "Fateh", "Amirreza", "" ] ]
TITLE: Efficient and Accurate Pneumonia Detection Using a Novel Multi-Scale Transformer Approach ABSTRACT: Pneumonia, a prevalent respiratory infection, remains a leading cause of morbidity and mortality worldwide, particularly among vulnerable populations. Chest X-rays serve as a primary tool for pneumonia detec...
2408.06828
Jingzhi Bao
Jingzhi Bao, Guanying Chen, Shuguang Cui
PIR: Photometric Inverse Rendering with Shading Cues Modeling and Surface Reflectance Regularization
Accepted to 3DV 2025. Project page: https://jzbao03.site/projects/PIR/
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
This paper addresses the problem of inverse rendering from photometric images. Existing approaches for this problem suffer from the effects of self-shadows, inter-reflections, and lack of constraints on the surface reflectance, leading to inaccurate decomposition of reflectance and illumination due to the ill-posed n...
[ { "version": "v1", "created": "Tue, 13 Aug 2024 11:39:14 GMT" }, { "version": "v2", "created": "Wed, 29 Jan 2025 17:18:18 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 03:08:44 GMT" } ]
2025-04-09T00:00:00
[ [ "Bao", "Jingzhi", "" ], [ "Chen", "Guanying", "" ], [ "Cui", "Shuguang", "" ] ]
TITLE: PIR: Photometric Inverse Rendering with Shading Cues Modeling and Surface Reflectance Regularization ABSTRACT: This paper addresses the problem of inverse rendering from photometric images. Existing approaches for this problem suffer from the effects of self-shadows, inter-reflections, and lack of constrai...
2408.12598
Ziyu Tang
Ziyu Tang, Weicai Ye, Yifan Wang, Di Huang, Hujun Bao, Tong He, Guofeng Zhang
ND-SDF: Learning Normal Deflection Fields for High-Fidelity Indoor Reconstruction
null
null
null
null
cs.CV cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Neural implicit reconstruction via volume rendering has demonstrated its effectiveness in recovering dense 3D surfaces. However, it is non-trivial to simultaneously recover meticulous geometry and preserve smoothness across regions with differing characteristics. To address this issue, previous methods typically empl...
[ { "version": "v1", "created": "Thu, 22 Aug 2024 17:59:01 GMT" }, { "version": "v2", "created": "Thu, 26 Sep 2024 06:31:25 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 15:24:36 GMT" } ]
2025-04-09T00:00:00
[ [ "Tang", "Ziyu", "" ], [ "Ye", "Weicai", "" ], [ "Wang", "Yifan", "" ], [ "Huang", "Di", "" ], [ "Bao", "Hujun", "" ], [ "He", "Tong", "" ], [ "Zhang", "Guofeng", "" ] ]
TITLE: ND-SDF: Learning Normal Deflection Fields for High-Fidelity Indoor Reconstruction ABSTRACT: Neural implicit reconstruction via volume rendering has demonstrated its effectiveness in recovering dense 3D surfaces. However, it is non-trivial to simultaneously recover meticulous geometry and preserve smoothnes...
2408.13378
Yoshitaka Inoue
Yoshitaka Inoue, Tianci Song, Xinling Wang, Augustin Luna, Tianfan Fu
DrugAgent: Multi-Agent Large Language Model-Based Reasoning for Drug-Target Interaction Prediction
15 pages, 1 figure
null
null
null
cs.AI cs.CL cs.IR cs.LG q-bio.QM
http://creativecommons.org/licenses/by/4.0/
Advancements in large language models (LLMs) allow them to address diverse questions using human-like interfaces. Still, limitations in their training prevent them from answering accurately in scenarios that could benefit from multiple perspectives. Multi-agent systems allow the resolution of questions to enhance res...
[ { "version": "v1", "created": "Fri, 23 Aug 2024 21:24:59 GMT" }, { "version": "v2", "created": "Thu, 12 Sep 2024 16:06:37 GMT" }, { "version": "v3", "created": "Mon, 16 Sep 2024 22:13:30 GMT" }, { "version": "v4", "created": "Mon, 7 Apr 2025 19:32:55 GMT" } ]
2025-04-09T00:00:00
[ [ "Inoue", "Yoshitaka", "" ], [ "Song", "Tianci", "" ], [ "Wang", "Xinling", "" ], [ "Luna", "Augustin", "" ], [ "Fu", "Tianfan", "" ] ]
TITLE: DrugAgent: Multi-Agent Large Language Model-Based Reasoning for Drug-Target Interaction Prediction ABSTRACT: Advancements in large language models (LLMs) allow them to address diverse questions using human-like interfaces. Still, limitations in their training prevent them from answering accurately in scena...
2409.00134
Alexey Skrynnik
Anton Andreychuk, Konstantin Yakovlev, Aleksandr Panov, Alexey Skrynnik
MAPF-GPT: Imitation Learning for Multi-Agent Pathfinding at Scale
null
null
null
null
cs.MA cs.AI cs.LG
http://creativecommons.org/licenses/by/4.0/
Multi-agent pathfinding (MAPF) is a problem that generally requires finding collision-free paths for multiple agents in a shared environment. Solving MAPF optimally, even under restrictive assumptions, is NP-hard, yet efficient solutions for this problem are critical for numerous applications, such as automated wareh...
[ { "version": "v1", "created": "Thu, 29 Aug 2024 12:55:10 GMT" }, { "version": "v2", "created": "Thu, 12 Sep 2024 13:49:00 GMT" }, { "version": "v3", "created": "Wed, 25 Sep 2024 13:09:35 GMT" }, { "version": "v4", "created": "Tue, 11 Feb 2025 12:28:36 GMT" }, { "v...
2025-04-09T00:00:00
[ [ "Andreychuk", "Anton", "" ], [ "Yakovlev", "Konstantin", "" ], [ "Panov", "Aleksandr", "" ], [ "Skrynnik", "Alexey", "" ] ]
TITLE: MAPF-GPT: Imitation Learning for Multi-Agent Pathfinding at Scale ABSTRACT: Multi-agent pathfinding (MAPF) is a problem that generally requires finding collision-free paths for multiple agents in a shared environment. Solving MAPF optimally, even under restrictive assumptions, is NP-hard, yet efficient solut...
2409.13717
Yiheng Wu
Yiheng Wu, Roman Yangarber, Xian Mao
DiVA-DocRE: A Discriminative and Voice-Aware Paradigm for Document-Level Relation Extraction
After internal discussions among the co-authors, we have decided to withdraw the manuscript due to a change in research direction and a lack of unanimous agreement to proceed with publication at this time
null
null
null
cs.CL cs.AI cs.IR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The remarkable capabilities of Large Language Models (LLMs) in text comprehension and generation have revolutionized Information Extraction (IE). One such advancement is in Document-level Relation Triplet Extraction (DocRTE), a critical task in information systems that aims to extract entities and their semantic rela...
[ { "version": "v1", "created": "Sat, 7 Sep 2024 18:47:38 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 10:43:00 GMT" } ]
2025-04-09T00:00:00
[ [ "Wu", "Yiheng", "" ], [ "Yangarber", "Roman", "" ], [ "Mao", "Xian", "" ] ]
TITLE: DiVA-DocRE: A Discriminative and Voice-Aware Paradigm for Document-Level Relation Extraction ABSTRACT: The remarkable capabilities of Large Language Models (LLMs) in text comprehension and generation have revolutionized Information Extraction (IE). One such advancement is in Document-level Relation Triplet...
2409.16681
Kun Zhou
Kun Zhou, You Zhang, Shengkui Zhao, Hao Wang, Zexu Pan, Dianwen Ng, Chong Zhang, Chongjia Ni, Yukun Ma, Trung Hieu Nguyen, Jia Qi Yip, Bin Ma
Emotional Dimension Control in Language Model-Based Text-to-Speech: Spanning a Broad Spectrum of Human Emotions
null
null
null
null
eess.AS cs.CL cs.SD
http://creativecommons.org/licenses/by/4.0/
Current emotional text-to-speech systems face challenges in conveying the full spectrum of human emotions, largely due to the inherent complexity of human emotions and the limited range of emotional labels in existing speech datasets. To address these limitations, this paper introduces a TTS framework that provides f...
[ { "version": "v1", "created": "Wed, 25 Sep 2024 07:16:16 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 08:08:08 GMT" } ]
2025-04-09T00:00:00
[ [ "Zhou", "Kun", "" ], [ "Zhang", "You", "" ], [ "Zhao", "Shengkui", "" ], [ "Wang", "Hao", "" ], [ "Pan", "Zexu", "" ], [ "Ng", "Dianwen", "" ], [ "Zhang", "Chong", "" ], [ "Ni", "Chongjia", ...
TITLE: Emotional Dimension Control in Language Model-Based Text-to-Speech: Spanning a Broad Spectrum of Human Emotions ABSTRACT: Current emotional text-to-speech systems face challenges in conveying the full spectrum of human emotions, largely due to the inherent complexity of human emotions and the limited range...
2410.05454
Ayesha Vermani
Ayesha Vermani, Josue Nassar, Hyungju Jeon, Matthew Dowling, Il Memming Park
Meta-Dynamical State Space Models for Integrative Neural Data Analysis
null
null
null
null
stat.ML cs.LG q-bio.NC
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Learning shared structure across environments facilitates rapid learning and adaptive behavior in neural systems. This has been widely demonstrated and applied in machine learning to train models that are capable of generalizing to novel settings. However, there has been limited work exploiting the shared structure i...
[ { "version": "v1", "created": "Mon, 7 Oct 2024 19:35:49 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 21:44:06 GMT" } ]
2025-04-09T00:00:00
[ [ "Vermani", "Ayesha", "" ], [ "Nassar", "Josue", "" ], [ "Jeon", "Hyungju", "" ], [ "Dowling", "Matthew", "" ], [ "Park", "Il Memming", "" ] ]
TITLE: Meta-Dynamical State Space Models for Integrative Neural Data Analysis ABSTRACT: Learning shared structure across environments facilitates rapid learning and adaptive behavior in neural systems. This has been widely demonstrated and applied in machine learning to train models that are capable of generalizing...
2410.08527
Yangyi Chen
Yangyi Chen, Binxuan Huang, Yifan Gao, Zhengyang Wang, Jingfeng Yang, Heng Ji
Scaling Laws for Predicting Downstream Performance in LLMs
Accepted to TMLR
null
null
null
cs.CL cs.AI cs.LG
http://creativecommons.org/licenses/by-nc-sa/4.0/
Precise estimation of downstream performance in large language models (LLMs) prior to training is essential for guiding their development process. Scaling laws analysis utilizes the statistics of a series of significantly smaller sampling language models (LMs) to predict the performance of the target LLM. For downstr...
[ { "version": "v1", "created": "Fri, 11 Oct 2024 04:57:48 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 21:47:09 GMT" } ]
2025-04-09T00:00:00
[ [ "Chen", "Yangyi", "" ], [ "Huang", "Binxuan", "" ], [ "Gao", "Yifan", "" ], [ "Wang", "Zhengyang", "" ], [ "Yang", "Jingfeng", "" ], [ "Ji", "Heng", "" ] ]
TITLE: Scaling Laws for Predicting Downstream Performance in LLMs ABSTRACT: Precise estimation of downstream performance in large language models (LLMs) prior to training is essential for guiding their development process. Scaling laws analysis utilizes the statistics of a series of significantly smaller sampling l...
2410.12779
Xingzhi Sun
Xingzhi Sun, Danqi Liao, Kincaid MacDonald, Yanlei Zhang, Chen Liu, Guillaume Huguet, Guy Wolf, Ian Adelstein, Tim G. J. Rudner, Smita Krishnaswamy
Geometry-Aware Generative Autoencoders for Warped Riemannian Metric Learning and Generative Modeling on Data Manifolds
Published in Proceedings of the 28th International Conference on Artificial Intelligence and Statistics (AISTATS 2025)
null
null
null
cs.LG math.DG stat.ML
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Rapid growth of high-dimensional datasets in fields such as single-cell RNA sequencing and spatial genomics has led to unprecedented opportunities for scientific discovery, but it also presents unique computational and statistical challenges. Traditional methods struggle with geometry-aware data generation, interpola...
[ { "version": "v1", "created": "Wed, 16 Oct 2024 17:53:26 GMT" }, { "version": "v2", "created": "Fri, 18 Oct 2024 18:27:10 GMT" }, { "version": "v3", "created": "Sat, 25 Jan 2025 16:39:26 GMT" }, { "version": "v4", "created": "Mon, 7 Apr 2025 19:30:58 GMT" } ]
2025-04-09T00:00:00
[ [ "Sun", "Xingzhi", "" ], [ "Liao", "Danqi", "" ], [ "MacDonald", "Kincaid", "" ], [ "Zhang", "Yanlei", "" ], [ "Liu", "Chen", "" ], [ "Huguet", "Guillaume", "" ], [ "Wolf", "Guy", "" ], [ "Adelstein"...
TITLE: Geometry-Aware Generative Autoencoders for Warped Riemannian Metric Learning and Generative Modeling on Data Manifolds ABSTRACT: Rapid growth of high-dimensional datasets in fields such as single-cell RNA sequencing and spatial genomics has led to unprecedented opportunities for scientific discovery, but i...
2410.16520
Naba Rizvi
Naba Rizvi, Harper Strickland, Daniel Gitelman, Tristan Cooper, Alexis Morales-Flores, Michael Golden, Aekta Kallepalli, Akshat Alurkar, Haaset Owens, Saleha Ahmedi, Isha Khirwadkar, Imani Munyaka, Nedjma Ousidhoum
AUTALIC: A Dataset for Anti-AUTistic Ableist Language In Context
9 pages, 5 figures, 7 tables
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by/4.0/
As our understanding of autism and ableism continues to increase, so does our understanding of ableist language towards autistic people. Such language poses a significant challenge in NLP research due to its subtle and context-dependent nature. Yet, detecting anti-autistic ableist language remains underexplored, with...
[ { "version": "v1", "created": "Mon, 21 Oct 2024 21:21:29 GMT" }, { "version": "v2", "created": "Thu, 21 Nov 2024 16:43:06 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 17:08:26 GMT" } ]
2025-04-09T00:00:00
[ [ "Rizvi", "Naba", "" ], [ "Strickland", "Harper", "" ], [ "Gitelman", "Daniel", "" ], [ "Cooper", "Tristan", "" ], [ "Morales-Flores", "Alexis", "" ], [ "Golden", "Michael", "" ], [ "Kallepalli", "Aekta", ""...
TITLE: AUTALIC: A Dataset for Anti-AUTistic Ableist Language In Context ABSTRACT: As our understanding of autism and ableism continues to increase, so does our understanding of ableist language towards autistic people. Such language poses a significant challenge in NLP research due to its subtle and context-depende...
2410.17875
Guangyuan Shi
Guangyuan Shi, Zexin Lu, Xiaoyu Dong, Wenlong Zhang, Xuanyu Zhang, Yujie Feng, Xiao-Ming Wu
Understanding Layer Significance in LLM Alignment
null
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by-nc-sa/4.0/
Aligning large language models (LLMs) through supervised fine-tuning is essential for tailoring them to specific applications. Recent studies suggest that alignment primarily adjusts a model's presentation style rather than its foundational knowledge, indicating that only certain components of the model are significa...
[ { "version": "v1", "created": "Wed, 23 Oct 2024 13:47:05 GMT" }, { "version": "v2", "created": "Fri, 20 Dec 2024 19:24:24 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 09:44:28 GMT" } ]
2025-04-09T00:00:00
[ [ "Shi", "Guangyuan", "" ], [ "Lu", "Zexin", "" ], [ "Dong", "Xiaoyu", "" ], [ "Zhang", "Wenlong", "" ], [ "Zhang", "Xuanyu", "" ], [ "Feng", "Yujie", "" ], [ "Wu", "Xiao-Ming", "" ] ]
TITLE: Understanding Layer Significance in LLM Alignment ABSTRACT: Aligning large language models (LLMs) through supervised fine-tuning is essential for tailoring them to specific applications. Recent studies suggest that alignment primarily adjusts a model's presentation style rather than its foundational knowledg...
2410.18358
Henrik Ebel
Henrik Ebel, Jan van Delden, Timo L\"uddecke, Aditya Borse, Rutwik Gulakala, Marcus Stoffel, Manish Yadav, Merten Stender, Leon Schindler, Kristin Miriam de Payrebrune, Maximilian Raff, C. David Remy, Benedict R\"oder, Rohit Raj, Tobias Rentschler, Alexander Tismer, Stefan Riedelbauch, Peter Eberhard
Data Publishing in Mechanics and Dynamics: Challenges, Guidelines, and Examples from Engineering Design
25 pages, 10 figures
DCE 6 (2025) e23
10.1017/dce.2025.13
null
cs.CY cs.AI cs.CE cs.ET cs.SY eess.SY
http://creativecommons.org/licenses/by/4.0/
Data-based methods have gained increasing importance in engineering, especially but not only driven by successes with deep artificial neural networks. Success stories are prevalent, e.g., in areas such as data-driven modeling, control and automation, as well as surrogate modeling for accelerated simulation. Beyond en...
[ { "version": "v1", "created": "Mon, 7 Oct 2024 18:26:05 GMT" }, { "version": "v2", "created": "Fri, 20 Dec 2024 12:58:09 GMT" } ]
2025-04-09T00:00:00
[ [ "Ebel", "Henrik", "" ], [ "van Delden", "Jan", "" ], [ "Lüddecke", "Timo", "" ], [ "Borse", "Aditya", "" ], [ "Gulakala", "Rutwik", "" ], [ "Stoffel", "Marcus", "" ], [ "Yadav", "Manish", "" ], [ "S...
TITLE: Data Publishing in Mechanics and Dynamics: Challenges, Guidelines, and Examples from Engineering Design ABSTRACT: Data-based methods have gained increasing importance in engineering, especially but not only driven by successes with deep artificial neural networks. Success stories are prevalent, e.g., in ar...
2410.19426
Daniel Galperin
Daniel Galperin, Ullrich K\"othe
Analyzing Generative Models by Manifold Entropic Metrics
Camera-ready version: accepted at AISTATS 2025
null
null
null
cs.LG stat.ML
http://creativecommons.org/licenses/by/4.0/
Good generative models should not only synthesize high quality data, but also utilize interpretable representations that aid human understanding of their behavior. However, it is difficult to measure objectively if and to what degree desirable properties of disentangled representations have been achieved. Inspired by...
[ { "version": "v1", "created": "Fri, 25 Oct 2024 09:35:00 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 15:47:53 GMT" } ]
2025-04-09T00:00:00
[ [ "Galperin", "Daniel", "" ], [ "Köthe", "Ullrich", "" ] ]
TITLE: Analyzing Generative Models by Manifold Entropic Metrics ABSTRACT: Good generative models should not only synthesize high quality data, but also utilize interpretable representations that aid human understanding of their behavior. However, it is difficult to measure objectively if and to what degree desirabl...
2411.02540
Mateusz Cedro
Mateusz Cedro, David Martens
GraphXAIN: Narratives to Explain Graph Neural Networks
19 pages, 9 figures, 2 tables
World Conference on Explainable Artificial Intelligence 2025
null
null
cs.LG cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Graph Neural Networks (GNNs) are a powerful technique for machine learning on graph-structured data, yet they pose challenges in interpretability. Existing GNN explanation methods usually yield technical outputs, such as subgraphs and feature importance scores, that are difficult for non-data scientists to understand...
[ { "version": "v1", "created": "Mon, 4 Nov 2024 19:21:06 GMT" }, { "version": "v2", "created": "Fri, 8 Nov 2024 08:29:10 GMT" }, { "version": "v3", "created": "Wed, 12 Feb 2025 15:14:01 GMT" } ]
2025-04-09T00:00:00
[ [ "Cedro", "Mateusz", "" ], [ "Martens", "David", "" ] ]
TITLE: GraphXAIN: Narratives to Explain Graph Neural Networks ABSTRACT: Graph Neural Networks (GNNs) are a powerful technique for machine learning on graph-structured data, yet they pose challenges in interpretability. Existing GNN explanation methods usually yield technical outputs, such as subgraphs and feature i...
2411.04794
Yuxin Zuo
Yuxin Zuo, Wenxuan Jiang, Wenxuan Liu, Zixuan Li, Long Bai, Hanbin Wang, Yutao Zeng, Xiaolong Jin, Jiafeng Guo, Xueqi Cheng
KnowCoder-X: Boosting Multilingual Information Extraction via Code
26 pages, 3 figures
null
null
null
cs.CL cs.AI cs.LG
http://creativecommons.org/licenses/by-nc-sa/4.0/
Empirical evidence indicates that LLMs exhibit spontaneous cross-lingual alignment. However, although LLMs show promising cross-lingual alignment in IE, a significant imbalance across languages persists, highlighting an underlying deficiency. To address this, we propose KnowCoder-X, a powerful code LLM with advanced ...
[ { "version": "v1", "created": "Thu, 7 Nov 2024 15:36:05 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 16:16:30 GMT" } ]
2025-04-09T00:00:00
[ [ "Zuo", "Yuxin", "" ], [ "Jiang", "Wenxuan", "" ], [ "Liu", "Wenxuan", "" ], [ "Li", "Zixuan", "" ], [ "Bai", "Long", "" ], [ "Wang", "Hanbin", "" ], [ "Zeng", "Yutao", "" ], [ "Jin", "Xiaolong",...
TITLE: KnowCoder-X: Boosting Multilingual Information Extraction via Code ABSTRACT: Empirical evidence indicates that LLMs exhibit spontaneous cross-lingual alignment. However, although LLMs show promising cross-lingual alignment in IE, a significant imbalance across languages persists, highlighting an underlying d...
2411.08872
Sadjad Alikhani
Sadjad Alikhani, Gouranga Charan, and Ahmed Alkhateeb
Large Wireless Model (LWM): A Foundation Model for Wireless Channels
The LWM model and relevant scripts are available on the LWM website: https://lwm-wireless.net/
null
null
null
cs.IT eess.SP math.IT
http://creativecommons.org/licenses/by-nc-sa/4.0/
This paper presents Large Wireless Model (LWM) -- the world's first foundation model for wireless channels. Designed as a task-agnostic model, LWM generates universal, rich, contextualized channel embeddings (features) that potentially enhance performance across a wide range of downstream tasks in wireless communicat...
[ { "version": "v1", "created": "Wed, 13 Nov 2024 18:51:10 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 19:49:37 GMT" } ]
2025-04-09T00:00:00
[ [ "Alikhani", "Sadjad", "" ], [ "Charan", "Gouranga", "" ], [ "Alkhateeb", "Ahmed", "" ] ]
TITLE: Large Wireless Model (LWM): A Foundation Model for Wireless Channels ABSTRACT: This paper presents Large Wireless Model (LWM) -- the world's first foundation model for wireless channels. Designed as a task-agnostic model, LWM generates universal, rich, contextualized channel embeddings (features) that potent...
2411.13951
Lucas Correia
Lucas Correia, Jan-Christoph Goos, Thomas B\"ack, Anna V. Kononova
PATH: A Discrete-sequence Dataset for Evaluating Online Unsupervised Anomaly Detection Approaches for Multivariate Time Series
Submitted to the Big Data Research journal
null
null
null
cs.LG cs.AI cs.CE cs.SY eess.SY
http://creativecommons.org/licenses/by/4.0/
Benchmarking anomaly detection approaches for multivariate time series is a challenging task due to a lack of high-quality datasets. Current publicly available datasets are too small, not diverse and feature trivial anomalies, which hinders measurable progress in this research area. We propose a solution: a diverse, ...
[ { "version": "v1", "created": "Thu, 21 Nov 2024 09:03:12 GMT" }, { "version": "v2", "created": "Mon, 25 Nov 2024 14:24:57 GMT" }, { "version": "v3", "created": "Wed, 15 Jan 2025 17:16:22 GMT" }, { "version": "v4", "created": "Tue, 8 Apr 2025 15:26:49 GMT" } ]
2025-04-09T00:00:00
[ [ "Correia", "Lucas", "" ], [ "Goos", "Jan-Christoph", "" ], [ "Bäck", "Thomas", "" ], [ "Kononova", "Anna V.", "" ] ]
TITLE: PATH: A Discrete-sequence Dataset for Evaluating Online Unsupervised Anomaly Detection Approaches for Multivariate Time Series ABSTRACT: Benchmarking anomaly detection approaches for multivariate time series is a challenging task due to a lack of high-quality datasets. Current publicly available datasets a...
2411.16199
Haojie Zheng
Shuchen Weng, Haojie Zheng, Peixuan Zhang, Yuchen Hong, Han Jiang, Si Li, Boxin Shi
VIRES: Video Instance Repainting via Sketch and Text Guided Generation
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
We introduce VIRES, a video instance repainting method with sketch and text guidance, enabling video instance repainting, replacement, generation, and removal. Existing approaches struggle with temporal consistency and accurate alignment with the provided sketch sequence. VIRES leverages the generative priors of text...
[ { "version": "v1", "created": "Mon, 25 Nov 2024 08:55:41 GMT" }, { "version": "v2", "created": "Tue, 26 Nov 2024 11:43:01 GMT" }, { "version": "v3", "created": "Fri, 28 Feb 2025 08:57:48 GMT" }, { "version": "v4", "created": "Tue, 4 Mar 2025 05:28:29 GMT" }, { "ve...
2025-04-09T00:00:00
[ [ "Weng", "Shuchen", "" ], [ "Zheng", "Haojie", "" ], [ "Zhang", "Peixuan", "" ], [ "Hong", "Yuchen", "" ], [ "Jiang", "Han", "" ], [ "Li", "Si", "" ], [ "Shi", "Boxin", "" ] ]
TITLE: VIRES: Video Instance Repainting via Sketch and Text Guided Generation ABSTRACT: We introduce VIRES, a video instance repainting method with sketch and text guidance, enabling video instance repainting, replacement, generation, and removal. Existing approaches struggle with temporal consistency and accurate ...
2411.16260
Fu-Chieh Chang
Fu-Chieh Chang, You-Chen Lin, Pei-Yuan Wu
Unraveling Arithmetic in Large Language Models: The Role of Algebraic Structures
null
ICLR 2025 Workshop on Reasoning and Planning for Large Language Models
null
null
cs.LG cs.CL
http://creativecommons.org/licenses/by/4.0/
The reasoning abilities of large language models (LLMs) have improved with chain-of-thought (CoT) prompting, allowing models to solve complex tasks stepwise. However, training CoT capabilities requires detailed reasoning data, which is often scarce. The self-taught reasoner (STaR) framework addresses this by using re...
[ { "version": "v1", "created": "Mon, 25 Nov 2024 10:23:11 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 15:19:23 GMT" } ]
2025-04-09T00:00:00
[ [ "Chang", "Fu-Chieh", "" ], [ "Lin", "You-Chen", "" ], [ "Wu", "Pei-Yuan", "" ] ]
TITLE: Unraveling Arithmetic in Large Language Models: The Role of Algebraic Structures ABSTRACT: The reasoning abilities of large language models (LLMs) have improved with chain-of-thought (CoT) prompting, allowing models to solve complex tasks stepwise. However, training CoT capabilities requires detailed reaso...
2411.16310
Jaime Corsetti
Jaime Corsetti, Francesco Giuliari, Alice Fasoli, Davide Boscaini, Fabio Poiesi
Functionality understanding and segmentation in 3D scenes
CVPR 2025 Highlight. Camera ready version. 20 pages, 12 figures, 7 tables
null
null
null
cs.CV
http://creativecommons.org/licenses/by/4.0/
Understanding functionalities in 3D scenes involves interpreting natural language descriptions to locate functional interactive objects, such as handles and buttons, in a 3D environment. Functionality understanding is highly challenging, as it requires both world knowledge to interpret language and spatial perception...
[ { "version": "v1", "created": "Mon, 25 Nov 2024 11:57:48 GMT" }, { "version": "v2", "created": "Tue, 26 Nov 2024 16:45:22 GMT" }, { "version": "v3", "created": "Wed, 4 Dec 2024 15:12:06 GMT" }, { "version": "v4", "created": "Tue, 8 Apr 2025 08:30:11 GMT" } ]
2025-04-09T00:00:00
[ [ "Corsetti", "Jaime", "" ], [ "Giuliari", "Francesco", "" ], [ "Fasoli", "Alice", "" ], [ "Boscaini", "Davide", "" ], [ "Poiesi", "Fabio", "" ] ]
TITLE: Functionality understanding and segmentation in 3D scenes ABSTRACT: Understanding functionalities in 3D scenes involves interpreting natural language descriptions to locate functional interactive objects, such as handles and buttons, in a 3D environment. Functionality understanding is highly challenging, as ...
2411.17191
Naoki Matsumura
Naoki Matsumura, Yuta Yoshimoto, Tamio Yamazaki, Tomohito Amano, Tomoyuki Noda, Naoki Ebata, Takatoshi Kasano and Yasufumi Sakai
Generator of Neural Network Potential for Molecular Dynamics: Constructing Robust and Accurate Potentials with Active Learning for Nanosecond-scale Simulations
null
null
10.1021/acs.jctc.4c01613
null
cond-mat.mtrl-sci physics.comp-ph
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Neural network potentials (NNPs) enable large-scale molecular dynamics (MD) simulations of systems containing >10,000 atoms with the accuracy comparable to ab initio methods and play a crucial role in material studies. Although NNPs are valuable for short-duration MD simulations, maintaining the stability of long-dur...
[ { "version": "v1", "created": "Tue, 26 Nov 2024 08:03:13 GMT" }, { "version": "v2", "created": "Wed, 19 Mar 2025 07:20:57 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 07:53:26 GMT" } ]
2025-04-09T00:00:00
[ [ "Matsumura", "Naoki", "" ], [ "Yoshimoto", "Yuta", "" ], [ "Yamazaki", "Tamio", "" ], [ "Amano", "Tomohito", "" ], [ "Noda", "Tomoyuki", "" ], [ "Ebata", "Naoki", "" ], [ "Kasano", "Takatoshi", "" ], [ ...
TITLE: Generator of Neural Network Potential for Molecular Dynamics: Constructing Robust and Accurate Potentials with Active Learning for Nanosecond-scale Simulations ABSTRACT: Neural network potentials (NNPs) enable large-scale molecular dynamics (MD) simulations of systems containing >10,000 atoms with the ac...
2412.06206
Nan Zhang
Nan Zhang, Prafulla Kumar Choubey, Alexander Fabbri, Gabriel Bernadett-Shapiro, Rui Zhang, Prasenjit Mitra, Caiming Xiong, Chien-Sheng Wu
SiReRAG: Indexing Similar and Related Information for Multihop Reasoning
ICLR 2025
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by/4.0/
Indexing is an important step towards strong performance in retrieval-augmented generation (RAG) systems. However, existing methods organize data based on either semantic similarity (similarity) or related information (relatedness), but do not cover both perspectives comprehensively. Our analysis reveals that modelin...
[ { "version": "v1", "created": "Mon, 9 Dec 2024 04:56:43 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 19:47:16 GMT" } ]
2025-04-09T00:00:00
[ [ "Zhang", "Nan", "" ], [ "Choubey", "Prafulla Kumar", "" ], [ "Fabbri", "Alexander", "" ], [ "Bernadett-Shapiro", "Gabriel", "" ], [ "Zhang", "Rui", "" ], [ "Mitra", "Prasenjit", "" ], [ "Xiong", "Caiming", ...
TITLE: SiReRAG: Indexing Similar and Related Information for Multihop Reasoning ABSTRACT: Indexing is an important step towards strong performance in retrieval-augmented generation (RAG) systems. However, existing methods organize data based on either semantic similarity (similarity) or related information (related...
2412.06717
Sahil Sethi
Sahil Sethi, Sai Reddy, Mansi Sakarvadia, Jordan Serotte, Darlington Nwaudo, Nicholas Maassen, Lewis Shi
Toward Non-Invasive Diagnosis of Bankart Lesions with Deep Learning
Accepted for presentation at SPIE Medical Imaging 2025: Computer-Aided Diagnosis. The manuscript is expected to appear in the conference proceedings
null
10.1117/12.3046251
null
cs.CV cs.AI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Bankart lesions, or anterior-inferior glenoid labral tears, are diagnostically challenging on standard MRIs due to their subtle imaging features-often necessitating invasive MRI arthrograms (MRAs). This study develops deep learning (DL) models to detect Bankart lesions on both standard MRIs and MRAs, aiming to improv...
[ { "version": "v1", "created": "Mon, 9 Dec 2024 18:04:27 GMT" } ]
2025-04-09T00:00:00
[ [ "Sethi", "Sahil", "" ], [ "Reddy", "Sai", "" ], [ "Sakarvadia", "Mansi", "" ], [ "Serotte", "Jordan", "" ], [ "Nwaudo", "Darlington", "" ], [ "Maassen", "Nicholas", "" ], [ "Shi", "Lewis", "" ] ]
TITLE: Toward Non-Invasive Diagnosis of Bankart Lesions with Deep Learning ABSTRACT: Bankart lesions, or anterior-inferior glenoid labral tears, are diagnostically challenging on standard MRIs due to their subtle imaging features-often necessitating invasive MRI arthrograms (MRAs). This study develops deep learning...
2412.06947
Bardia Nadimi
Bardia Nadimi and Ghali Omar Boutaib and Hao Zheng
PyraNet: A Multi-Layered Hierarchical Dataset for Verilog
null
null
null
null
cs.AR cs.AI cs.LG cs.PL
http://creativecommons.org/licenses/by-nc-sa/4.0/
Recently, there has been a growing interest in leveraging Large Language Models for Verilog code generation. However, the current quality of the generated Verilog code remains suboptimal. This is largely due to the absence of well-defined, well-organized datasets with high-quality samples, as well as a lack of innova...
[ { "version": "v1", "created": "Mon, 9 Dec 2024 19:45:54 GMT" }, { "version": "v2", "created": "Fri, 27 Dec 2024 01:07:02 GMT" }, { "version": "v3", "created": "Mon, 7 Apr 2025 21:58:26 GMT" } ]
2025-04-09T00:00:00
[ [ "Nadimi", "Bardia", "" ], [ "Boutaib", "Ghali Omar", "" ], [ "Zheng", "Hao", "" ] ]
TITLE: PyraNet: A Multi-Layered Hierarchical Dataset for Verilog ABSTRACT: Recently, there has been a growing interest in leveraging Large Language Models for Verilog code generation. However, the current quality of the generated Verilog code remains suboptimal. This is largely due to the absence of well-defined, w...
2412.07456
Ben Steinfurth
Jonas Schulte-Sasse, Ben Steinfurth and Julien Weiss
Automatic extraction of wall streamlines from oil-flow visualizations using a convolutional neural network
null
Exp. Fluids 66 (2025)
10.1007/s00348-025-04016-x
null
physics.flu-dyn
http://creativecommons.org/licenses/by/4.0/
Oil-flow visualizations represent a simple means to reveal time-averaged wall streamline patterns. Yet, the evaluation of such images can be a time-consuming process and is subjective to human perception. In this study, we present a fast and robust method to obtain quantitative insight based on qualitative oil-flow v...
[ { "version": "v1", "created": "Tue, 10 Dec 2024 12:21:44 GMT" } ]
2025-04-09T00:00:00
[ [ "Schulte-Sasse", "Jonas", "" ], [ "Steinfurth", "Ben", "" ], [ "Weiss", "Julien", "" ] ]
TITLE: Automatic extraction of wall streamlines from oil-flow visualizations using a convolutional neural network ABSTRACT: Oil-flow visualizations represent a simple means to reveal time-averaged wall streamline patterns. Yet, the evaluation of such images can be a time-consuming process and is subjective to hum...
2412.08307
Shijian Wang
Shijian Wang, Linxin Song, Jieyu Zhang, Ryotaro Shimizu, Jiarui Jin, Ao Luo, Yuan Lu, Li Yao, Cunjian Chen, Julian McAuley, Wentao Zhang, Hanqian Wu
Investigating the Scaling Effect of Instruction Templates for Training Multimodal Language Model
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Current multimodal language model (MLM) training approaches overlook the influence of instruction templates. Previous research deals with this problem by leveraging hand-crafted or model-generated templates, failing to investigate the scaling effect of instruction templates on MLM training. In this work, we propose a...
[ { "version": "v1", "created": "Wed, 11 Dec 2024 11:39:42 GMT" }, { "version": "v2", "created": "Sun, 6 Apr 2025 14:45:49 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 08:30:30 GMT" } ]
2025-04-09T00:00:00
[ [ "Wang", "Shijian", "" ], [ "Song", "Linxin", "" ], [ "Zhang", "Jieyu", "" ], [ "Shimizu", "Ryotaro", "" ], [ "Jin", "Jiarui", "" ], [ "Luo", "Ao", "" ], [ "Lu", "Yuan", "" ], [ "Yao", "Li", ...
TITLE: Investigating the Scaling Effect of Instruction Templates for Training Multimodal Language Model ABSTRACT: Current multimodal language model (MLM) training approaches overlook the influence of instruction templates. Previous research deals with this problem by leveraging hand-crafted or model-generated tem...
2412.08755
Kyle Stein
Kyle Stein, Andrew Arash Mahyari, Guillermo Francia, Eman El-Sheikh
Proactive Adversarial Defense: Harnessing Prompt Tuning in Vision-Language Models to Detect Unseen Backdoored Images
null
null
null
null
cs.CV cs.AI cs.CR cs.LG
http://creativecommons.org/licenses/by/4.0/
Backdoor attacks pose a critical threat by embedding hidden triggers into inputs, causing models to misclassify them into target labels. While extensive research has focused on mitigating these attacks in object recognition models through weight fine-tuning, much less attention has been given to detecting backdoored ...
[ { "version": "v1", "created": "Wed, 11 Dec 2024 19:54:14 GMT" }, { "version": "v2", "created": "Thu, 9 Jan 2025 19:15:20 GMT" }, { "version": "v3", "created": "Fri, 14 Mar 2025 19:24:34 GMT" }, { "version": "v4", "created": "Mon, 7 Apr 2025 18:01:26 GMT" } ]
2025-04-09T00:00:00
[ [ "Stein", "Kyle", "" ], [ "Mahyari", "Andrew Arash", "" ], [ "Francia", "Guillermo", "" ], [ "El-Sheikh", "Eman", "" ] ]
TITLE: Proactive Adversarial Defense: Harnessing Prompt Tuning in Vision-Language Models to Detect Unseen Backdoored Images ABSTRACT: Backdoor attacks pose a critical threat by embedding hidden triggers into inputs, causing models to misclassify them into target labels. While extensive research has focused on mit...
2412.11530
Junda Cheng
Junda Cheng, Zhipeng Cai, Zhaoxing Zhang, Wei Yin, Matthias Muller, Michael Paulitsch, Xin Yang
RoMeO: Robust Metric Visual Odometry
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Visual odometry (VO) aims to estimate camera poses from visual inputs -- a fundamental building block for many applications such as VR/AR and robotics. This work focuses on monocular RGB VO where the input is a monocular RGB video without IMU or 3D sensors. Existing approaches lack robustness under this challenging s...
[ { "version": "v1", "created": "Mon, 16 Dec 2024 08:08:35 GMT" }, { "version": "v2", "created": "Thu, 19 Dec 2024 06:32:22 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 13:16:35 GMT" } ]
2025-04-09T00:00:00
[ [ "Cheng", "Junda", "" ], [ "Cai", "Zhipeng", "" ], [ "Zhang", "Zhaoxing", "" ], [ "Yin", "Wei", "" ], [ "Muller", "Matthias", "" ], [ "Paulitsch", "Michael", "" ], [ "Yang", "Xin", "" ] ]
TITLE: RoMeO: Robust Metric Visual Odometry ABSTRACT: Visual odometry (VO) aims to estimate camera poses from visual inputs -- a fundamental building block for many applications such as VR/AR and robotics. This work focuses on monocular RGB VO where the input is a monocular RGB video without IMU or 3D sensors. Exis...
2412.17867
Ziming Guo
Ziming Guo, Chao Ma, Yinggang Sun, Tiancheng Zhao, Guangyao Wang, Hai Huang
Evaluating and Enhancing LLMs for Multi-turn Text-to-SQL with Multiple Question Types
International Joint Conference on Neural Networks 2025 (IJCNN 2025)
null
null
null
cs.CL cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Recent advancements in large language models (LLMs) have significantly advanced text-to-SQL systems. However, most LLM-based methods often narrowly focus on SQL generation, neglecting the complexities of real-world conversational queries. This oversight can lead to unreliable responses, particularly for ambiguous que...
[ { "version": "v1", "created": "Sat, 21 Dec 2024 10:13:45 GMT" }, { "version": "v2", "created": "Thu, 3 Apr 2025 07:13:30 GMT" }, { "version": "v3", "created": "Sun, 6 Apr 2025 09:47:45 GMT" }, { "version": "v4", "created": "Tue, 8 Apr 2025 02:23:17 GMT" } ]
2025-04-09T00:00:00
[ [ "Guo", "Ziming", "" ], [ "Ma", "Chao", "" ], [ "Sun", "Yinggang", "" ], [ "Zhao", "Tiancheng", "" ], [ "Wang", "Guangyao", "" ], [ "Huang", "Hai", "" ] ]
TITLE: Evaluating and Enhancing LLMs for Multi-turn Text-to-SQL with Multiple Question Types ABSTRACT: Recent advancements in large language models (LLMs) have significantly advanced text-to-SQL systems. However, most LLM-based methods often narrowly focus on SQL generation, neglecting the complexities of real-wo...
2501.00952
Maxim Ziatdinov
Sarah I. Allec, Maxim Ziatdinov
Active and transfer learning with partially Bayesian neural networks for materials and chemicals
Minor revisions
null
null
null
cond-mat.dis-nn cond-mat.mtrl-sci physics.data-an
http://creativecommons.org/licenses/by/4.0/
Active learning, an iterative process of selecting the most informative data points for exploration, is crucial for efficient characterization of materials and chemicals property space. Neural networks excel at predicting these properties but lack the uncertainty quantification needed for active learning-driven explo...
[ { "version": "v1", "created": "Wed, 1 Jan 2025 20:48:26 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 20:33:33 GMT" } ]
2025-04-09T00:00:00
[ [ "Allec", "Sarah I.", "" ], [ "Ziatdinov", "Maxim", "" ] ]
TITLE: Active and transfer learning with partially Bayesian neural networks for materials and chemicals ABSTRACT: Active learning, an iterative process of selecting the most informative data points for exploration, is crucial for efficient characterization of materials and chemicals property space. Neural network...
2501.04671
Charles Corbi\`ere
Charles Corbi\`ere, Simon Roburin, Syrielle Montariol, Antoine Bosselut and Alexandre Alahi
Retrieval-Based Interleaved Visual Chain-of-Thought in Real-World Driving Scenarios
Project page: https://vita-epfl.github.io/DrivingVQA
null
null
null
cs.CV cs.AI
http://creativecommons.org/licenses/by/4.0/
While chain-of-thought (CoT) prompting improves reasoning in large language models, its effectiveness in vision-language models (VLMs) remains limited due to over-reliance on textual cues and memorized knowledge. To investigate the visual reasoning capabilities of VLMs in complex real-world scenarios, we introduce Dr...
[ { "version": "v1", "created": "Wed, 8 Jan 2025 18:31:16 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 17:09:59 GMT" } ]
2025-04-09T00:00:00
[ [ "Corbière", "Charles", "" ], [ "Roburin", "Simon", "" ], [ "Montariol", "Syrielle", "" ], [ "Bosselut", "Antoine", "" ], [ "Alahi", "Alexandre", "" ] ]
TITLE: Retrieval-Based Interleaved Visual Chain-of-Thought in Real-World Driving Scenarios ABSTRACT: While chain-of-thought (CoT) prompting improves reasoning in large language models, its effectiveness in vision-language models (VLMs) remains limited due to over-reliance on textual cues and memorized knowledge. ...
2501.05446
Yifan Yu
Yifan Yu, Shaohui Liu, R\'emi Pautrat, Marc Pollefeys, Viktor Larsson
Relative Pose Estimation through Affine Corrections of Monocular Depth Priors
CVPR 2025 (Highlight)
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Monocular depth estimation (MDE) models have undergone significant advancements over recent years. Many MDE models aim to predict affine-invariant relative depth from monocular images, while recent developments in large-scale training and vision foundation models enable reasonable estimation of metric (absolute) dept...
[ { "version": "v1", "created": "Thu, 9 Jan 2025 18:58:30 GMT" }, { "version": "v2", "created": "Mon, 24 Mar 2025 17:14:43 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 03:59:21 GMT" } ]
2025-04-09T00:00:00
[ [ "Yu", "Yifan", "" ], [ "Liu", "Shaohui", "" ], [ "Pautrat", "Rémi", "" ], [ "Pollefeys", "Marc", "" ], [ "Larsson", "Viktor", "" ] ]
TITLE: Relative Pose Estimation through Affine Corrections of Monocular Depth Priors ABSTRACT: Monocular depth estimation (MDE) models have undergone significant advancements over recent years. Many MDE models aim to predict affine-invariant relative depth from monocular images, while recent developments in large...
2501.09333
Wei-Lun Chao
Arpita Chowdhury, Dipanjyoti Paul, Zheda Mai, Jianyang Gu, Ziheng Zhang, Kazi Sajeed Mehrab, Elizabeth G. Campolongo, Daniel Rubenstein, Charles V. Stewart, Anuj Karpatne, Tanya Berger-Wolf, Yu Su, Wei-Lun Chao
Prompt-CAM: Making Vision Transformers Interpretable for Fine-Grained Analysis
Accepted by CVPR 2025 Main Conference
null
null
null
cs.CV cs.AI
http://creativecommons.org/licenses/by-nc-sa/4.0/
We present a simple approach to make pre-trained Vision Transformers (ViTs) interpretable for fine-grained analysis, aiming to identify and localize the traits that distinguish visually similar categories, such as bird species. Pre-trained ViTs, such as DINO, have demonstrated remarkable capabilities in extracting lo...
[ { "version": "v1", "created": "Thu, 16 Jan 2025 07:07:41 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 18:03:40 GMT" } ]
2025-04-09T00:00:00
[ [ "Chowdhury", "Arpita", "" ], [ "Paul", "Dipanjyoti", "" ], [ "Mai", "Zheda", "" ], [ "Gu", "Jianyang", "" ], [ "Zhang", "Ziheng", "" ], [ "Mehrab", "Kazi Sajeed", "" ], [ "Campolongo", "Elizabeth G.", "" ...
TITLE: Prompt-CAM: Making Vision Transformers Interpretable for Fine-Grained Analysis ABSTRACT: We present a simple approach to make pre-trained Vision Transformers (ViTs) interpretable for fine-grained analysis, aiming to identify and localize the traits that distinguish visually similar categories, such as bird...
2501.11014
Ken Enda
Ken Enda, Yoshitaka Oda, Zen-ichi Tanei, Kenichi Satoh, Hiroaki Motegi, Terasaka Shunsuke, Shigeru Yamaguchi, Takahiro Ogawa, Wang Lei, Masumi Tsuda and Shinya Tanaka
Transfer Learning Strategies for Pathological Foundation Models: A Systematic Evaluation in Brain Tumor Classification
25 pages, 7 figures
null
null
null
eess.IV cs.CV
http://creativecommons.org/licenses/by/4.0/
Foundation models pretrained on large-scale pathology datasets have shown promising results across various diagnostic tasks. Here, we present a systematic evaluation of transfer learning strategies for brain tumor classification using these models. We analyzed 254 cases comprising five major tumor types: glioblastoma...
[ { "version": "v1", "created": "Sun, 19 Jan 2025 11:18:34 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 01:49:45 GMT" } ]
2025-04-09T00:00:00
[ [ "Enda", "Ken", "" ], [ "Oda", "Yoshitaka", "" ], [ "Tanei", "Zen-ichi", "" ], [ "Satoh", "Kenichi", "" ], [ "Motegi", "Hiroaki", "" ], [ "Shunsuke", "Terasaka", "" ], [ "Yamaguchi", "Shigeru", "" ], [ ...
TITLE: Transfer Learning Strategies for Pathological Foundation Models: A Systematic Evaluation in Brain Tumor Classification ABSTRACT: Foundation models pretrained on large-scale pathology datasets have shown promising results across various diagnostic tasks. Here, we present a systematic evaluation of transfer ...
2501.17848
Fabricio Olivetti de Franca
Fabricio Olivetti de Franca and Gabriel Kronberger
Improving Genetic Programming for Symbolic Regression with Equality Graphs
10 pages, 5 figures, 4 tables. In Genetic and Evolutionary Computation Conference (GECCO 25)
null
10.1145/3712256.3726383
null
cs.LG
http://creativecommons.org/licenses/by/4.0/
The search for symbolic regression models with genetic programming (GP) has a tendency of revisiting expressions in their original or equivalent forms. Repeatedly evaluating equivalent expressions is inefficient, as it does not immediately lead to better solutions. However, evolutionary algorithms require diversity a...
[ { "version": "v1", "created": "Wed, 29 Jan 2025 18:49:34 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 16:48:10 GMT" } ]
2025-04-09T00:00:00
[ [ "de Franca", "Fabricio Olivetti", "" ], [ "Kronberger", "Gabriel", "" ] ]
TITLE: Improving Genetic Programming for Symbolic Regression with Equality Graphs ABSTRACT: The search for symbolic regression models with genetic programming (GP) has a tendency of revisiting expressions in their original or equivalent forms. Repeatedly evaluating equivalent expressions is inefficient, as it doe...
2502.03251
Li Sun
Li Sun, Zhenhao Huang, Suyang Zhou, Qiqi Wan, Hao Peng, Philip Yu
RiemannGFM: Learning a Graph Foundation Model from Riemannian Geometry
Accepted by WWW 2025 (Oral)
null
null
null
cs.LG
http://creativecommons.org/licenses/by-nc-sa/4.0/
The foundation model has heralded a new era in artificial intelligence, pretraining a single model to offer cross-domain transferability on different datasets. Graph neural networks excel at learning graph data, the omnipresent non-Euclidean structure, but often lack the generalization capacity. Hence, graph foundati...
[ { "version": "v1", "created": "Wed, 5 Feb 2025 15:06:09 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 07:04:29 GMT" } ]
2025-04-09T00:00:00
[ [ "Sun", "Li", "" ], [ "Huang", "Zhenhao", "" ], [ "Zhou", "Suyang", "" ], [ "Wan", "Qiqi", "" ], [ "Peng", "Hao", "" ], [ "Yu", "Philip", "" ] ]
TITLE: RiemannGFM: Learning a Graph Foundation Model from Riemannian Geometry ABSTRACT: The foundation model has heralded a new era in artificial intelligence, pretraining a single model to offer cross-domain transferability on different datasets. Graph neural networks excel at learning graph data, the omnipresent ...
2502.04760
Rui Wang
Rui Wang
Graph Federated Learning Based Proactive Content Caching in Edge Computing
null
null
null
null
cs.LG cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
With the rapid growth of mobile data traffic and the increasing prevalence of video streaming, proactive content caching in edge computing has become crucial for reducing latency and alleviating network congestion. However, traditional caching strategies such as FIFO, LRU, and LFU fail to effectively predict future c...
[ { "version": "v1", "created": "Fri, 7 Feb 2025 08:48:06 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 12:46:45 GMT" } ]
2025-04-09T00:00:00
[ [ "Wang", "Rui", "" ] ]
TITLE: Graph Federated Learning Based Proactive Content Caching in Edge Computing ABSTRACT: With the rapid growth of mobile data traffic and the increasing prevalence of video streaming, proactive content caching in edge computing has become crucial for reducing latency and alleviating network congestion. However...
2502.07847
Behraj Khan
Behraj Khan, Rizwan Qureshi, Nouman Muhammad Durrani, Tahir Syed
Confidence-calibrated covariate shift correction for few-shot classification in Vision-Language Models
null
null
null
null
cs.CV cs.LG
http://creativecommons.org/licenses/by-nc-sa/4.0/
Since the establishment of vision-language foundation models as the new mainstay in low-shot vision classification tasks, the question of domain generalization arising from insufficient target data is assuming more importance. This scarcity challenge induces sampling bias and amplifies model sensitivity to variations...
[ { "version": "v1", "created": "Tue, 11 Feb 2025 10:10:15 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 07:54:30 GMT" } ]
2025-04-09T00:00:00
[ [ "Khan", "Behraj", "" ], [ "Qureshi", "Rizwan", "" ], [ "Durrani", "Nouman Muhammad", "" ], [ "Syed", "Tahir", "" ] ]
TITLE: Confidence-calibrated covariate shift correction for few-shot classification in Vision-Language Models ABSTRACT: Since the establishment of vision-language foundation models as the new mainstay in low-shot vision classification tasks, the question of domain generalization arising from insufficient target d...
2502.11007
Liangqi Yuan
Liangqi Yuan and Dong-Jun Han and Shiqiang Wang and Christopher G. Brinton
Local-Cloud Inference Offloading for LLMs in Multi-Modal, Multi-Task, Multi-Dialogue Settings
null
null
null
null
cs.LG cs.DC
http://creativecommons.org/licenses/by/4.0/
Compared to traditional machine learning models, recent large language models (LLMs) can exhibit multi-task-solving capabilities through multiple dialogues and multi-modal data sources. These unique characteristics of LLMs, together with their large model size, make their deployment more challenging. Specifically, (i...
[ { "version": "v1", "created": "Sun, 16 Feb 2025 06:18:28 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 18:49:28 GMT" } ]
2025-04-09T00:00:00
[ [ "Yuan", "Liangqi", "" ], [ "Han", "Dong-Jun", "" ], [ "Wang", "Shiqiang", "" ], [ "Brinton", "Christopher G.", "" ] ]
TITLE: Local-Cloud Inference Offloading for LLMs in Multi-Modal, Multi-Task, Multi-Dialogue Settings ABSTRACT: Compared to traditional machine learning models, recent large language models (LLMs) can exhibit multi-task-solving capabilities through multiple dialogues and multi-modal data sources. These unique char...
2502.14270
Rajeshwari Mistri
Nachiket Kapure, Harsh Joshi, Rajeshwari Mistri, Parul Kumari, Manasi Mali, Seema Purohit, Neha Sharma, Mrityunjoy Panday, Chittaranjan S. Yajnik
Predicting Fetal Birthweight from High Dimensional Data using Advanced Machine Learning
null
null
null
null
cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Birth weight serves as a fundamental indicator of neonatal health, closely linked to both early medical interventions and long-term developmental risks. Traditional predictive models, often constrained by limited feature selection and incomplete datasets, struggle to achieve overlooking complex maternal and fetal int...
[ { "version": "v1", "created": "Thu, 20 Feb 2025 05:17:39 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 07:54:17 GMT" } ]
2025-04-09T00:00:00
[ [ "Kapure", "Nachiket", "" ], [ "Joshi", "Harsh", "" ], [ "Mistri", "Rajeshwari", "" ], [ "Kumari", "Parul", "" ], [ "Mali", "Manasi", "" ], [ "Purohit", "Seema", "" ], [ "Sharma", "Neha", "" ], [ "Pa...
TITLE: Predicting Fetal Birthweight from High Dimensional Data using Advanced Machine Learning ABSTRACT: Birth weight serves as a fundamental indicator of neonatal health, closely linked to both early medical interventions and long-term developmental risks. Traditional predictive models, often constrained by limi...
2502.19363
Ru Peng
Ru Peng, Kexin Yang, Yawen Zeng, Junyang Lin, Dayiheng Liu, Junbo Zhao
DataMan: Data Manager for Pre-training Large Language Models
ICLR2025 paper
null
null
null
cs.CL cs.AI
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The performance emergence of large language models (LLMs) driven by data scaling laws makes the selection of pre-training data increasingly important. However, existing methods rely on limited heuristics and human intuition, lacking comprehensive and clear guidelines. To address this, we are inspired by ``reverse thi...
[ { "version": "v1", "created": "Wed, 26 Feb 2025 18:01:19 GMT" }, { "version": "v2", "created": "Thu, 13 Mar 2025 15:42:07 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 03:21:10 GMT" } ]
2025-04-09T00:00:00
[ [ "Peng", "Ru", "" ], [ "Yang", "Kexin", "" ], [ "Zeng", "Yawen", "" ], [ "Lin", "Junyang", "" ], [ "Liu", "Dayiheng", "" ], [ "Zhao", "Junbo", "" ] ]
TITLE: DataMan: Data Manager for Pre-training Large Language Models ABSTRACT: The performance emergence of large language models (LLMs) driven by data scaling laws makes the selection of pre-training data increasingly important. However, existing methods rely on limited heuristics and human intuition, lacking compr...
2502.19679
Linzhuo Li
Linzhuo li
Old Experience Helps: Leveraging Survey Methodology to Improve AI Text Annotation Reliability in Social Sciences
7 figures
null
null
null
cs.DL cs.HC
http://creativecommons.org/licenses/by/4.0/
This paper introduces a framework for assessing the reliability of Large Language Model (LLM) text annotations in social science research by adapting established survey methodology principles. Drawing parallels between survey respondent behavior and LLM outputs, the study implements three key interventions: option ra...
[ { "version": "v1", "created": "Thu, 27 Feb 2025 01:42:10 GMT" }, { "version": "v2", "created": "Thu, 13 Mar 2025 03:06:47 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 06:48:04 GMT" } ]
2025-04-09T00:00:00
[ [ "li", "Linzhuo", "" ] ]
TITLE: Old Experience Helps: Leveraging Survey Methodology to Improve AI Text Annotation Reliability in Social Sciences ABSTRACT: This paper introduces a framework for assessing the reliability of Large Language Model (LLM) text annotations in social science research by adapting established survey methodology pri...
2502.21024
Abdelrahman E.M. Abdallah
Abdelrahman Abdallah, Bhawna Piryani, Jonas Wallat, Avishek Anand, Adam Jatowt
TempRetriever: Fusion-based Temporal Dense Passage Retrieval for Time-Sensitive Questions
null
null
null
null
cs.IR cs.CL
http://creativecommons.org/licenses/by/4.0/
Temporal awareness is crucial in many information retrieval tasks, particularly in scenarios where the relevance of documents depends on their alignment with the query's temporal context. Traditional approaches such as BM25 and Dense Passage Retrieval (DPR) focus on lexical or semantic similarity but tend to neglect ...
[ { "version": "v1", "created": "Fri, 28 Feb 2025 13:06:25 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 13:11:58 GMT" } ]
2025-04-09T00:00:00
[ [ "Abdallah", "Abdelrahman", "" ], [ "Piryani", "Bhawna", "" ], [ "Wallat", "Jonas", "" ], [ "Anand", "Avishek", "" ], [ "Jatowt", "Adam", "" ] ]
TITLE: TempRetriever: Fusion-based Temporal Dense Passage Retrieval for Time-Sensitive Questions ABSTRACT: Temporal awareness is crucial in many information retrieval tasks, particularly in scenarios where the relevance of documents depends on their alignment with the query's temporal context. Traditional approac...
2503.05050
Melkamu Mersha
Melkamu Abay Mersha, Mesay Gemeda Yigezu, Hassan Shakil, Ali K. AlShami, Sanghyun Byun, Jugal Kalita
A Unified Framework with Novel Metrics for Evaluating the Effectiveness of XAI Techniques in LLMs
arXiv admin note: substantial text overlap with arXiv:2501.15374
null
null
null
cs.CL cs.AI cs.LG
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
The increasing complexity of LLMs presents significant challenges to their transparency and interpretability, necessitating the use of eXplainable AI (XAI) techniques to enhance trustworthiness and usability. This study introduces a comprehensive evaluation framework with four novel metrics for assessing the effectiv...
[ { "version": "v1", "created": "Thu, 6 Mar 2025 23:59:50 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 20:37:11 GMT" } ]
2025-04-09T00:00:00
[ [ "Mersha", "Melkamu Abay", "" ], [ "Yigezu", "Mesay Gemeda", "" ], [ "Shakil", "Hassan", "" ], [ "AlShami", "Ali K.", "" ], [ "Byun", "Sanghyun", "" ], [ "Kalita", "Jugal", "" ] ]
TITLE: A Unified Framework with Novel Metrics for Evaluating the Effectiveness of XAI Techniques in LLMs ABSTRACT: The increasing complexity of LLMs presents significant challenges to their transparency and interpretability, necessitating the use of eXplainable AI (XAI) techniques to enhance trustworthiness and u...
2503.05725
Kim Duc Tran
T.Q.D. Pham, K.D. Tran, Khanh T. P. Nguyen, X.V. Tran, L. K\"oehl, and K.P. Tran
A new framework for prognostics in decentralized industries: Enhancing fairness, security, and transparency through Blockchain and Federated Learning
null
null
null
null
cs.CY cs.AI
http://creativecommons.org/licenses/by/4.0/
As global industries transition towards Industry 5.0 predictive maintenance PM remains crucial for cost effective operations resilience and minimizing downtime in increasingly smart manufacturing environments In this chapter we explore how the integration of Federated Learning FL and blockchain BC technologies enhanc...
[ { "version": "v1", "created": "Mon, 17 Feb 2025 20:28:40 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 16:53:33 GMT" } ]
2025-04-09T00:00:00
[ [ "Pham", "T. Q. D.", "" ], [ "Tran", "K. D.", "" ], [ "Nguyen", "Khanh T. P.", "" ], [ "Tran", "X. V.", "" ], [ "Köehl", "L.", "" ], [ "Tran", "K. P.", "" ] ]
TITLE: A new framework for prognostics in decentralized industries: Enhancing fairness, security, and transparency through Blockchain and Federated Learning ABSTRACT: As global industries transition towards Industry 5.0 predictive maintenance PM remains crucial for cost effective operations resilience and minim...
2503.07378
Yusuke Hashimoto
Yusuke Hashimoto, Xue Jia, Hao Li, Takaaki Tomai
A Materials Map Integrating Experimental and Computational Data via Graph-Based Machine Learning for Enhanced Materials Discovery
null
null
null
null
cond-mat.mtrl-sci cs.LG
http://creativecommons.org/licenses/by/4.0/
Materials informatics (MI), emerging from the integration of materials science and data science, is expected to significantly accelerate material development and discovery. The data used in MI are derived from both computational and experimental studies; however, their integration remains challenging. In our previous...
[ { "version": "v1", "created": "Mon, 10 Mar 2025 14:31:34 GMT" }, { "version": "v2", "created": "Tue, 11 Mar 2025 06:31:52 GMT" }, { "version": "v3", "created": "Thu, 13 Mar 2025 10:04:14 GMT" }, { "version": "v4", "created": "Tue, 18 Mar 2025 04:43:10 GMT" }, { "v...
2025-04-09T00:00:00
[ [ "Hashimoto", "Yusuke", "" ], [ "Jia", "Xue", "" ], [ "Li", "Hao", "" ], [ "Tomai", "Takaaki", "" ] ]
TITLE: A Materials Map Integrating Experimental and Computational Data via Graph-Based Machine Learning for Enhanced Materials Discovery ABSTRACT: Materials informatics (MI), emerging from the integration of materials science and data science, is expected to significantly accelerate material development and disco...
2503.08111
Jianhui Wang
Jianhui Wang, Zhifei Yang, Yangfan He, Huixiong Zhang, Yuxuan Chen, Jingwei Huang
MaRI: Material Retrieval Integration across Domains
null
null
null
null
cs.CV
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Accurate material retrieval is critical for creating realistic 3D assets. Existing methods rely on datasets that capture shape-invariant and lighting-varied representations of materials, which are scarce and face challenges due to limited diversity and inadequate real-world generalization. Most current approaches ado...
[ { "version": "v1", "created": "Tue, 11 Mar 2025 07:23:11 GMT" }, { "version": "v2", "created": "Thu, 3 Apr 2025 07:30:21 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 08:53:57 GMT" } ]
2025-04-09T00:00:00
[ [ "Wang", "Jianhui", "" ], [ "Yang", "Zhifei", "" ], [ "He", "Yangfan", "" ], [ "Zhang", "Huixiong", "" ], [ "Chen", "Yuxuan", "" ], [ "Huang", "Jingwei", "" ] ]
TITLE: MaRI: Material Retrieval Integration across Domains ABSTRACT: Accurate material retrieval is critical for creating realistic 3D assets. Existing methods rely on datasets that capture shape-invariant and lighting-varied representations of materials, which are scarce and face challenges due to limited diversit...
2503.09516
Bowen Jin
Bowen Jin, Hansi Zeng, Zhenrui Yue, Jinsung Yoon, Sercan Arik, Dong Wang, Hamed Zamani, Jiawei Han
Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning
31 pages
null
null
null
cs.CL cs.AI cs.IR
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Efficiently acquiring external knowledge and up-to-date information is essential for effective reasoning and text generation in large language models (LLMs). Prompting advanced LLMs with reasoning capabilities to use search engines during inference is often suboptimal, as the LLM might not fully possess the capabilit...
[ { "version": "v1", "created": "Wed, 12 Mar 2025 16:26:39 GMT" }, { "version": "v2", "created": "Wed, 19 Mar 2025 21:40:12 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 14:03:26 GMT" } ]
2025-04-09T00:00:00
[ [ "Jin", "Bowen", "" ], [ "Zeng", "Hansi", "" ], [ "Yue", "Zhenrui", "" ], [ "Yoon", "Jinsung", "" ], [ "Arik", "Sercan", "" ], [ "Wang", "Dong", "" ], [ "Zamani", "Hamed", "" ], [ "Han", "Jiawei"...
TITLE: Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning ABSTRACT: Efficiently acquiring external knowledge and up-to-date information is essential for effective reasoning and text generation in large language models (LLMs). Prompting advanced LLMs with reasoning capabilit...
2503.12763
Kewei Sui
Kewei Sui, Anindita Ghosh, Inwoo Hwang, Bing Zhou, Jian Wang, Chuan Guo
A Survey on Human Interaction Motion Generation
The repository listing relevant papers is accessible at: https://github.com/soraproducer/Awesome-Human-Interaction-Motion-Generation
null
null
null
cs.CV cs.LG
http://creativecommons.org/licenses/by/4.0/
Humans inhabit a world defined by interactions -- with other humans, objects, and environments. These interactive movements not only convey our relationships with our surroundings but also demonstrate how we perceive and communicate with the real world. Therefore, replicating these interaction behaviors in digital sy...
[ { "version": "v1", "created": "Mon, 17 Mar 2025 02:55:10 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 23:38:41 GMT" } ]
2025-04-09T00:00:00
[ [ "Sui", "Kewei", "" ], [ "Ghosh", "Anindita", "" ], [ "Hwang", "Inwoo", "" ], [ "Zhou", "Bing", "" ], [ "Wang", "Jian", "" ], [ "Guo", "Chuan", "" ] ]
TITLE: A Survey on Human Interaction Motion Generation ABSTRACT: Humans inhabit a world defined by interactions -- with other humans, objects, and environments. These interactive movements not only convey our relationships with our surroundings but also demonstrate how we perceive and communicate with the real worl...
2503.17486
Zhengqing Gao
Zhengqing Gao, Dongting Hu, Jia-Wang Bian, Huan Fu, Yan Li, Tongliang Liu, Mingming Gong, Kun Zhang
ProtoGS: Efficient and High-Quality Rendering with 3D Gaussian Prototypes
null
null
null
null
cs.CV cs.AI
http://creativecommons.org/licenses/by/4.0/
3D Gaussian Splatting (3DGS) has made significant strides in novel view synthesis but is limited by the substantial number of Gaussian primitives required, posing challenges for deployment on lightweight devices. Recent methods address this issue by compressing the storage size of densified Gaussians, yet fail to pre...
[ { "version": "v1", "created": "Fri, 21 Mar 2025 18:55:14 GMT" }, { "version": "v2", "created": "Tue, 25 Mar 2025 13:03:48 GMT" }, { "version": "v3", "created": "Tue, 8 Apr 2025 12:19:01 GMT" } ]
2025-04-09T00:00:00
[ [ "Gao", "Zhengqing", "" ], [ "Hu", "Dongting", "" ], [ "Bian", "Jia-Wang", "" ], [ "Fu", "Huan", "" ], [ "Li", "Yan", "" ], [ "Liu", "Tongliang", "" ], [ "Gong", "Mingming", "" ], [ "Zhang", "Kun...
TITLE: ProtoGS: Efficient and High-Quality Rendering with 3D Gaussian Prototypes ABSTRACT: 3D Gaussian Splatting (3DGS) has made significant strides in novel view synthesis but is limited by the substantial number of Gaussian primitives required, posing challenges for deployment on lightweight devices. Recent met...
2503.22926
Zikang Yuan
Zikang Yuan, Ruiye Ming, Chengwei Zhao, Yonghao Tan, Pingcheng Dong, Hongcheng Luo, Yuzhong Jiao, Xin Yang and Kwang-Ting Cheng
SR-LIO++: Efficient LiDAR-Inertial Odometry and Quantized Mapping with Sweep Reconstruction
10 pages, 12 figures
null
null
null
cs.RO
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Addressing the inherent low acquisition frequency limitation of 3D LiDAR to achieve high-frequency output has become a critical research focus in the LiDAR-Inertial Odometry (LIO) domain. To ensure real-time performance, frequency-enhanced LIO systems must process each sweep within significantly reduced timeframe, wh...
[ { "version": "v1", "created": "Sat, 29 Mar 2025 01:06:54 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 05:27:15 GMT" } ]
2025-04-09T00:00:00
[ [ "Yuan", "Zikang", "" ], [ "Ming", "Ruiye", "" ], [ "Zhao", "Chengwei", "" ], [ "Tan", "Yonghao", "" ], [ "Dong", "Pingcheng", "" ], [ "Luo", "Hongcheng", "" ], [ "Jiao", "Yuzhong", "" ], [ "Yang", ...
TITLE: SR-LIO++: Efficient LiDAR-Inertial Odometry and Quantized Mapping with Sweep Reconstruction ABSTRACT: Addressing the inherent low acquisition frequency limitation of 3D LiDAR to achieve high-frequency output has become a critical research focus in the LiDAR-Inertial Odometry (LIO) domain. To ensure real-ti...
2504.00597
Jirui Qi
Jirui Qi, Raquel Fern\'andez, Arianna Bisazza
On the Consistency of Multilingual Context Utilization in Retrieval-Augmented Generation
Under review at COLM2025. All codes and data are released at https://github.com/Betswish/mRAG-Context-Consistency
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by/4.0/
Retrieval-augmented generation (RAG) with large language models (LLMs) has demonstrated strong performance in multilingual question-answering (QA) tasks by leveraging relevant passages retrieved from corpora. In multilingual RAG (mRAG), the retrieved passages can be written in languages other than that of the query e...
[ { "version": "v1", "created": "Tue, 1 Apr 2025 09:55:23 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 12:40:23 GMT" } ]
2025-04-09T00:00:00
[ [ "Qi", "Jirui", "" ], [ "Fernández", "Raquel", "" ], [ "Bisazza", "Arianna", "" ] ]
TITLE: On the Consistency of Multilingual Context Utilization in Retrieval-Augmented Generation ABSTRACT: Retrieval-augmented generation (RAG) with large language models (LLMs) has demonstrated strong performance in multilingual question-answering (QA) tasks by leveraging relevant passages retrieved from corpora....
2504.01698
Yilong Lu
Yi-Long Lu, Chunhui Zhang, Jiajun Song, Lifeng Fan, Wei Wang
ToM-RL: Reinforcement Learning Unlocks Theory of Mind in Small LLMs
null
null
null
null
cs.CL cs.AI
http://creativecommons.org/licenses/by/4.0/
Recent advancements in rule-based reinforcement learning (RL), applied during the post-training phase of large language models (LLMs), have significantly enhanced their capabilities in structured reasoning tasks such as mathematics and logical inference. However, the effectiveness of RL in social reasoning, particula...
[ { "version": "v1", "created": "Wed, 2 Apr 2025 12:58:42 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 03:58:20 GMT" } ]
2025-04-09T00:00:00
[ [ "Lu", "Yi-Long", "" ], [ "Zhang", "Chunhui", "" ], [ "Song", "Jiajun", "" ], [ "Fan", "Lifeng", "" ], [ "Wang", "Wei", "" ] ]
TITLE: ToM-RL: Reinforcement Learning Unlocks Theory of Mind in Small LLMs ABSTRACT: Recent advancements in rule-based reinforcement learning (RL), applied during the post-training phase of large language models (LLMs), have significantly enhanced their capabilities in structured reasoning tasks such as mathematics...
2504.02010
Nan Zhang
Nan Zhang, Yusen Zhang, Prasenjit Mitra, Rui Zhang
When Reasoning Meets Compression: Benchmarking Compressed Large Reasoning Models on Complex Reasoning Tasks
null
null
null
null
cs.LG cs.AI
http://creativecommons.org/licenses/by/4.0/
Recent open-source large reasoning models (LRMs) exhibit strong performance on complex reasoning tasks, but their large parameter count makes them prohibitively expensive for individuals. The compression of large language models (LLMs) offers an effective solution to reduce cost of computational resources. However, s...
[ { "version": "v1", "created": "Wed, 2 Apr 2025 05:17:46 GMT" } ]
2025-04-09T00:00:00
[ [ "Zhang", "Nan", "" ], [ "Zhang", "Yusen", "" ], [ "Mitra", "Prasenjit", "" ], [ "Zhang", "Rui", "" ] ]
TITLE: When Reasoning Meets Compression: Benchmarking Compressed Large Reasoning Models on Complex Reasoning Tasks ABSTRACT: Recent open-source large reasoning models (LRMs) exhibit strong performance on complex reasoning tasks, but their large parameter count makes them prohibitively expensive for individuals. T...
2504.02329
Seif Mzoughi Msc
Seif Mzoughi, Ahmed Haj yahmed, Mohamed Elshafei, Foutse Khomh, Diego Elias Costa
Towards Assessing Deep Learning Test Input Generators
Accepted to EASE 2025
null
null
null
cs.LG cs.CV cs.SE
http://creativecommons.org/licenses/by/4.0/
Deep Learning (DL) systems are increasingly deployed in safety-critical applications, yet they remain vulnerable to robustness issues that can lead to significant failures. While numerous Test Input Generators (TIGs) have been developed to evaluate DL robustness, a comprehensive assessment of their effectiveness acro...
[ { "version": "v1", "created": "Thu, 3 Apr 2025 07:06:55 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 18:35:13 GMT" } ]
2025-04-09T00:00:00
[ [ "Mzoughi", "Seif", "" ], [ "yahmed", "Ahmed Haj", "" ], [ "Elshafei", "Mohamed", "" ], [ "Khomh", "Foutse", "" ], [ "Costa", "Diego Elias", "" ] ]
TITLE: Towards Assessing Deep Learning Test Input Generators ABSTRACT: Deep Learning (DL) systems are increasingly deployed in safety-critical applications, yet they remain vulnerable to robustness issues that can lead to significant failures. While numerous Test Input Generators (TIGs) have been developed to evalu...
2504.02971
Shaoyuan Xu Ph.D.
Binh M. Le, Shaoyuan Xu, Jinmiao Fu, Zhishen Huang, Moyan Li, Yanhui Guo, Hongdong Li, Sameera Ramasinghe, Bryan Wang
QID: Efficient Query-Informed ViTs in Data-Scarce Regimes for OCR-free Visual Document Understanding
8 pages, accepted by CVPR 2025 MULA
null
null
null
cs.CV cs.CL
http://creativecommons.org/licenses/by/4.0/
In Visual Document Understanding (VDU) tasks, fine-tuning a pre-trained Vision-Language Model (VLM) with new datasets often falls short in optimizing the vision encoder to identify query-specific regions in text-rich document images. Existing methods that directly inject queries into model layers by modifying the net...
[ { "version": "v1", "created": "Thu, 3 Apr 2025 18:47:16 GMT" }, { "version": "v2", "created": "Mon, 7 Apr 2025 17:58:44 GMT" } ]
2025-04-09T00:00:00
[ [ "Le", "Binh M.", "" ], [ "Xu", "Shaoyuan", "" ], [ "Fu", "Jinmiao", "" ], [ "Huang", "Zhishen", "" ], [ "Li", "Moyan", "" ], [ "Guo", "Yanhui", "" ], [ "Li", "Hongdong", "" ], [ "Ramasinghe", "S...
TITLE: QID: Efficient Query-Informed ViTs in Data-Scarce Regimes for OCR-free Visual Document Understanding ABSTRACT: In Visual Document Understanding (VDU) tasks, fine-tuning a pre-trained Vision-Language Model (VLM) with new datasets often falls short in optimizing the vision encoder to identify query-specific ...
2504.03809
Niclas Boehmer
Stanis{\l}aw Szufa, Niclas Boehmer, Robert Bredereck, Piotr Faliszewski, Rolf Niedermeier, Piotr Skowron, Arkadii Slinko, Nimrod Talmon
Drawing a Map of Elections
Journal article merging results from arxiv:2105.07815, arXiv:2407.11889 and Szufa et al., "Drawing a Map of Elections in the Space of Statistical Cultures", AAMAS '20
null
10.1016/j.artint.2025.104332
null
cs.MA cs.AI cs.GT
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Our main contribution is the introduction of the map of elections framework. A map of elections consists of three main elements: (1) a dataset of elections (i.e., collections of ordinal votes over given sets of candidates), (2) a way of measuring similarities between these elections, and (3) a representation of the e...
[ { "version": "v1", "created": "Fri, 4 Apr 2025 11:44:56 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 10:52:54 GMT" } ]
2025-04-09T00:00:00
[ [ "Szufa", "Stanisław", "" ], [ "Boehmer", "Niclas", "" ], [ "Bredereck", "Robert", "" ], [ "Faliszewski", "Piotr", "" ], [ "Niedermeier", "Rolf", "" ], [ "Skowron", "Piotr", "" ], [ "Slinko", "Arkadii", "" ...
TITLE: Drawing a Map of Elections ABSTRACT: Our main contribution is the introduction of the map of elections framework. A map of elections consists of three main elements: (1) a dataset of elections (i.e., collections of ordinal votes over given sets of candidates), (2) a way of measuring similarities between thes...
2504.03814
Grgur Kova\v{c}
Grgur Kova\v{c}, J\'er\'emy Perez, R\'emy Portelas, Peter Ford Dominey, Pierre-Yves Oudeyer
Recursive Training Loops in LLMs: How training data properties modulate distribution shift in generated data?
null
null
null
null
cs.LG cs.AI cs.CL
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Large language models (LLMs) are increasingly contributing to the creation of content on the Internet. This creates a feedback loop as subsequent generations of models will be trained on this generated, synthetic data. This phenomenon is receiving increasing interest, in particular because previous studies have shown...
[ { "version": "v1", "created": "Fri, 4 Apr 2025 14:41:41 GMT" }, { "version": "v2", "created": "Tue, 8 Apr 2025 08:45:26 GMT" } ]
2025-04-09T00:00:00
[ [ "Kovač", "Grgur", "" ], [ "Perez", "Jérémy", "" ], [ "Portelas", "Rémy", "" ], [ "Dominey", "Peter Ford", "" ], [ "Oudeyer", "Pierre-Yves", "" ] ]
TITLE: Recursive Training Loops in LLMs: How training data properties modulate distribution shift in generated data? ABSTRACT: Large language models (LLMs) are increasingly contributing to the creation of content on the Internet. This creates a feedback loop as subsequent generations of models will be trained on ...