大模型对齐与安全 - arXivDaily 专题

2410.00863 2024-10-02 cs.CL 57%

On the Implications of Verbose LLM Outputs: A Case Study in Translation Evaluation

Eleftheria Briakou, Zhongtao Liu, Colin Cherry, Markus Freitag

专题命中安全评测：safety(abstract)；分类 cs.CL

2410.00699 2024-10-02 cs.LG stat.ML 57%

Investigating the Impact of Model Complexity in Large Language Models

Jing Luo, Huiyuan Wang, Weiran Huang

专题命中安全评测：alignment(abstract)；分类 cs.LG

2410.00175 2024-10-02 cs.CL 57%

Adaptable Moral Stances of Large Language Models on Sexist Content: Implications for Society and Gender Discourse

Rongchen Guo, Isar Nejadgholi, Hillary Dawkins, Kathleen C. Fraser, Svetlana Kiritchenko

专题命中安全评测：safety(abstract)；分类 cs.CL

Comments To be published at EMNLP2024

2408.09839 2024-10-02 cs.CV cs.AI 57%

Segment-Anything Models Achieve Zero-shot Robustness in Autonomous Driving

Jun Yan, Pengyu Wang, Danni Wang, Weiquan Huang, Daniel Watzenig, Huilin Yin

专题命中安全评测：trustworthy(abstract)；分类 cs.AI

Comments Accepted to IAVVC 2024

2406.08958 2024-10-01 cs.LG 57%

An Unsupervised Approach to Achieve Supervised-Level Explainability in Healthcare Records

Joakim Edin, Maria Maistro, Lars Maaløe, Lasse Borgholt, Jakob D. Havtorn, Tuukka Ruotsalo

专题命中安全评测：safety(abstract)；分类 cs.LG

Comments Accepted to EMNLP 2024 Main

2309.13475 2024-10-01 cs.RO cs.CV cs.LG cs.SY eess.SY 57%

Detecting and Mitigating System-Level Anomalies of Vision-Based Controllers

Aryaman Gupta, Kaustav Chakraborty, Somil Bansal

专题命中安全评测：safety(abstract)；分类 cs.LG

Journal ref 2024/5/13 Conference 2024 IEEE International Conference on Robotics and Automation (ICRA) Pages 9953-9959 Publisher 2024 IEEE International Conference on Robotics and Automation (ICRA)

2409.17012 2024-09-26 cs.AI 57%

AI-Driven Risk-Aware Scheduling for Active Debris Removal Missions

Antoine Poupon, Hugo de Rohan Willner, Pierre Nikitits, Adam Abdin

专题命中安全评测：safety(abstract)；分类 cs.AI

2409.16899 2024-09-26 cs.RO cs.CL cs.HC 57%

Robotic Backchanneling in Online Conversation Facilitation: A Cross-Generational Study

Sota Kobuki, Katie Seaborn, Seiki Tokunaga, Kosuke Fukumori, Shun Hidaka, Kazuhiro Tamura, Koji Inoue, Tatsuya Kawahara, Mihoko Otake-Mastuura

专题命中安全评测：trustworthy(abstract)；分类 cs.CL

Comments Published at Proceedings of the 2023 32nd IEEE International Conference on Robot and Human Interactive Communication (RO-MAN 2023)

2409.15740 2024-09-25 cs.AI cs.CV cs.NI 57%

Real-Time Pedestrian Detection on IoT Edge Devices: A Lightweight Deep Learning Approach

Muhammad Dany Alfikri, Rafael Kaliski

专题命中安全评测：safety(abstract)；分类 cs.AI

Comments 10 pages, 3 tables, 12 figures, article submitted to IEEE for possible publication

2409.14907 2024-09-24 cs.CL 57%

Knowledge Planning in Large Language Models for Domain-Aligned Counseling Summarization

Aseem Srivastava, Smriti Joshi, Tanmoy Chakraborty, Md Shad Akhtar

专题命中安全评测：alignment(abstract)；分类 cs.CL

Comments Full paper accepted at EMNLP 2024 (main)

2409.13741 2024-09-24 cs.CL cs.IR 57%

Knowing When to Ask -- Bridging Large Language Models and Data

Prashanth Radhakrishnan, Jennifer Chen, Bo Xu, Prem Ramaswami, Hannah Pho, Adriana Olmos, James Manyika, R. V. Guha

专题命中安全评测：trustworthy(abstract)；分类 cs.CL

Comments 39 pages - 25 page paper, 14 page Appendix, 7 figures, 9 tables

2304.10392 2024-09-24 cs.CL 57%

CKBP v2: Better Annotation and Reasoning for Commonsense Knowledge Base Population

Tianqing Fang, Quyet V. Do, Zihao Zheng, Weiqi Wang, Sehyun Choi, Zhaowei Wang, Yangqiu Song

专题命中安全评测：alignment(abstract)；分类 cs.CL

2404.13940 2024-09-23 cs.CL 57%

A User-Centric Multi-Intent Benchmark for Evaluating Large Language Models

Jiayin Wang, Fengran Mo, Weizhi Ma, Peijie Sun, Min Zhang, Jian-Yun Nie

专题命中安全评测：alignment(abstract)；分类 cs.CL

2409.12171 2024-09-19 cs.CR cs.AI 57%

Semantic Interoperability on Blockchain by Generating Smart Contracts Based on Knowledge Graphs

William Van Woensel, Oshani Seneviratne

专题命中安全评测：trustworthy(abstract)；分类 cs.AI

2409.12076 2024-09-19 cs.LG 57%

Unsupervised Domain Adaptation Via Data Pruning

Andrea Napoli, Paul White

专题命中安全评测：alignment(abstract)；分类 cs.LG

2409.11929 2024-09-19 cs.LG 57%

An Explainable Machine Learning Approach to Traffic Accident Fatality Prediction

Md. Asif Khan Rifat, Ahmedul Kabir, Armana Sabiha Huq

专题命中安全评测：safety(abstract)；分类 cs.LG

Comments 10 Pages, 6 figures, 2 tables, 28th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2024)

2409.10883 2024-09-18 cs.CL 57%

CREAM: Comparison-Based Reference-Free ELO-Ranked Automatic Evaluation for Meeting Summarization

Ziwei Gong, Lin Ai, Harshsaiprasad Deshpande, Alexander Johnson, Emmy Phung, Zehui Wu, Ahmad Emami, Julia Hirschberg

专题命中安全评测：alignment(abstract)；分类 cs.CL

2406.14367 2024-09-17 cs.CV cs.AI 57%

PoseBench: Benchmarking the Robustness of Pose Estimation Models under Corruptions

Sihan Ma, Jing Zhang, Qiong Cao, Dacheng Tao

专题命中安全评测：safety(abstract)；分类 cs.AI

Comments Technical report. Project page: https://xymsh.github.io/PoseBench/

2409.08887 2024-09-16 cs.CV cs.CL 57%

Visual Language Tracking with Multi-modal Interaction: A Robust Benchmark

Xuchen Li, Shiyu Hu, Xiaokun Feng, Dailing Zhang, Meiqi Wu, Jing Zhang, Kaiqi Huang

专题命中安全评测：alignment(abstract)；分类 cs.CL

Comments Under Review

2409.07652 2024-09-13 stat.ML cs.LG math.ST stat.TH 57%

Gaussian Process Upper Confidence Bounds in Distributed Point Target Tracking over Wireless Sensor Networks

Xingchi Liu, Lyudmila Mihaylova, Jemin George, Tien Pham

专题命中安全评测：trustworthy(abstract)；分类 cs.LG

2406.11774 2024-09-13 cs.LG cs.SY eess.SY 57%

Optimal Transport-Assisted Risk-Sensitive Q-Learning

Zahra Shahrooei, Ali Baheri

专题命中安全评测：safety(abstract)；分类 cs.LG

2409.05482 2024-09-10 astro-ph.SR astro-ph.EP astro-ph.IM cs.LG 57%

Advancing Machine Learning for Stellar Activity and Exoplanet Period Rotation

Fatemeh Fazel Hesar, Bernard Foing, Ana M. Heras, Mojtaba Raouf, Victoria Foing, Shima Javanmardi, Fons J. Verbeek

专题命中安全评测：alignment(abstract)；分类 cs.LG

Comments 15 pages, 8 figures. Submitted for publication in A&A

2409.05076 2024-09-10 cs.CV cs.AI 57%

PIP: Detecting Adversarial Examples in Large Vision-Language Models via Attention Patterns of Irrelevant Probe Questions

Yudong Zhang, Ruobing Xie, Jiansheng Chen, Xingwu Sun, Yu Wang

专题命中安全评测：safety(abstract)；分类 cs.AI

Comments Accepted by ACM Multimedia 2024 BNI track (Oral)

2409.04808 2024-09-10 cs.AI 57%

HULLMI: Human vs LLM identification with explainability

Prathamesh Dinesh Joshi, Sahil Pocker, Raj Abhijit Dandekar, Rajat Dandekar, Sreedath Panat

专题命中安全评测：trustworthy(abstract)；分类 cs.AI

Comments 17 pages, 8 figures

2405.09176 2024-09-10 cs.LG cs.CR 57%

Cross-Input Certified Training for Universal Perturbations

Changming Xu, Gagandeep Singh

专题命中安全评测：trustworthy(abstract)；分类 cs.LG

Comments 23 pages, 6 figures, ECCV '24

2409.04693 2024-09-10 cs.AI 57%

MuAP: Multi-step Adaptive Prompt Learning for Vision-Language Model with Missing Modality

Ruiting Dai, Yuqiao Tan, Lisi Mo, Tao He, Ke Qin, Shuang Liang

专题命中安全评测：alignment(abstract)；分类 cs.AI

2409.03598 2024-09-06 cs.LG cs.CV 57%

A practical approach to evaluating the adversarial distance for machine learning classifiers

Georg Siedel, Ekagra Gupta, Andrey Morozov

专题命中安全评测：safety(abstract)；分类 cs.LG

Comments Accepted manuscript at International Mechanical Engineering Congress and Exposition IMECE2024

2409.01037 2024-09-04 cs.CL 57%

NYK-MS: A Well-annotated Multi-modal Metaphor and Sarcasm Understanding Benchmark on Cartoon-Caption Dataset

Ke Chang, Hao Li, Junzhao Zhang, Yunfang Wu

专题命中安全评测：alignment(abstract)；分类 cs.CL

Comments 13 pages, 6 figures

2408.12762 2024-09-04 cs.HC cs.AI 57%

Visual Verity in AI-Generated Imagery: Computational Metrics and Human-Centric Analysis

Memoona Aziz, Umair Rehman, Syed Ali Safi, Amir Zaib Abbasi

专题命中安全评测：alignment(abstract)；分类 cs.AI

2403.02270 2024-09-04 cs.CL 57%

FENICE: Factuality Evaluation of summarization based on Natural language Inference and Claim Extraction

Alessandro Scirè, Karim Ghonim, Roberto Navigli

专题命中安全评测：alignment(abstract)；分类 cs.CL

Comments ACL 2024 camera ready. Code and data at https://github.com/Babelscape/FENICE

1. 安全评测 9434 篇

On the Implications of Verbose LLM Outputs: A Case Study in Translation Evaluation

Investigating the Impact of Model Complexity in Large Language Models

Adaptable Moral Stances of Large Language Models on Sexist Content: Implications for Society and Gender Discourse

Segment-Anything Models Achieve Zero-shot Robustness in Autonomous Driving

An Unsupervised Approach to Achieve Supervised-Level Explainability in Healthcare Records

Detecting and Mitigating System-Level Anomalies of Vision-Based Controllers

AI-Driven Risk-Aware Scheduling for Active Debris Removal Missions

Robotic Backchanneling in Online Conversation Facilitation: A Cross-Generational Study

Real-Time Pedestrian Detection on IoT Edge Devices: A Lightweight Deep Learning Approach

Knowledge Planning in Large Language Models for Domain-Aligned Counseling Summarization

Knowing When to Ask -- Bridging Large Language Models and Data

CKBP v2: Better Annotation and Reasoning for Commonsense Knowledge Base Population

A User-Centric Multi-Intent Benchmark for Evaluating Large Language Models

Semantic Interoperability on Blockchain by Generating Smart Contracts Based on Knowledge Graphs

Unsupervised Domain Adaptation Via Data Pruning

An Explainable Machine Learning Approach to Traffic Accident Fatality Prediction

CREAM: Comparison-Based Reference-Free ELO-Ranked Automatic Evaluation for Meeting Summarization

PoseBench: Benchmarking the Robustness of Pose Estimation Models under Corruptions

Visual Language Tracking with Multi-modal Interaction: A Robust Benchmark

Gaussian Process Upper Confidence Bounds in Distributed Point Target Tracking over Wireless Sensor Networks

Optimal Transport-Assisted Risk-Sensitive Q-Learning

Advancing Machine Learning for Stellar Activity and Exoplanet Period Rotation

PIP: Detecting Adversarial Examples in Large Vision-Language Models via Attention Patterns of Irrelevant Probe Questions

HULLMI: Human vs LLM identification with explainability

Cross-Input Certified Training for Universal Perturbations

MuAP: Multi-step Adaptive Prompt Learning for Vision-Language Model with Missing Modality

A practical approach to evaluating the adversarial distance for machine learning classifiers

NYK-MS: A Well-annotated Multi-modal Metaphor and Sarcasm Understanding Benchmark on Cartoon-Caption Dataset

Visual Verity in AI-Generated Imagery: Computational Metrics and Human-Centric Analysis

FENICE: Factuality Evaluation of summarization based on Natural language Inference and Claim Extraction