大模型对齐与安全

ESA: Energy-Based Shot Assembly Optimization for Automatic Video Editing

Yaosen Chen, Wei Wang, Tianheng Zheng, Xuming Wen, Han Yang, Yanru Zhang

机构 * Sobey Media Intelligence Laboratory（索贝媒体智能实验室） ； University of Electronic Science and Technology of China（电子科学与技术大学） ； SiChuan University（四川大学） ； Qinghai Normal University（青海师范大学） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

AdCare-VLM: Towards a Unified and Pre-aligned Latent Representation for Healthcare Video Understanding

Md Asaduzzaman Jabin, Hanqi Jiang, Yiwei Li, Patrick Kaggwa, Eugene Douglass, Juliet N. Sekandi, Tianming Liu

机构 * University of Georgia（佐治亚大学） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

Comments 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Workshop: 7th International Workshop on Large Scale Holistic Video Understanding: Toward Video Foundation Models

Journal ref Neural Information Processing Systems (NeurIPS 2025)

GMAT: Grounded Multi-Agent Clinical Description Generation for Text Encoder in Vision-Language MIL for Whole Slide Image Classification

Ngoc Bui Lam Quang, Nam Le Nguyen Binh, Thanh-Huy Nguyen, Le Thien Phuc Nguyen, Quan Nguyen, Ulas Bagci

机构 * AI VIETNAM（AI越南） ； Carnegie Mellon University（卡内基梅隆大学） ； University of Wisconsin-Madison（威斯康星大学麦迪逊分校） ； PTIT ； Northwestern University（西北大学） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

Comments Acccepted in MICCAI Workshop 2025

Generalizable and Fast Surrogates: Model Predictive Control of Articulated Soft Robots using Physics-Informed Neural Networks

Tim-Lukas Habich, Aran Mohammad, Simon F. G. Ehlers, Martin Bensch, Thomas Seel, Moritz Schappler

机构 * Leibniz University Hannover（莱布尼茨汉诺威大学） 纠错

专题命中 其他安全 ：safety(abstract)；分类 cs.LG

Comments Accepted for publication in IEEE Transactions on Robotics (T-RO) 2025

DINO-Detect: A Simple yet Effective Framework for Blur-Robust AI-Generated Image Detection

Jialiang Shen, Jiyang Zheng, Yunqi Xue, Huajie Chen, Yu Yao, Hui Kang, Ruiqi Liu, Helin Gong, Yang Yang, Dadong Wang, Tongliang Liu

机构 * Sydney AI Center, The University of Sydney（悉尼人工智能中心，悉尼大学） ； CSIRO, Data61（澳大利亚联邦科学与工业研究组织、Data61） ； Shanghai Jiao Tong University（上海交通大学） ； City University of Macau（澳门城市大学） ； CASIA（中国科学院自动化研究所） 纠错

专题命中 其他安全 ：safety(abstract)；分类 cs.LG

Comments 12 pages, 5 figures

VSPO: Validating Semantic Pitfalls in Ontology via LLM-Based CQ Generation

Hyojun Choi, Seokju Hwang, Kyong-Ho Lee

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

Comments Accepted at AAAI 2026 oral

DenseAnnotate: Enabling Scalable Dense Caption Collection for Images and 3D Scenes via Spoken Descriptions

Xiaoyu Lin, Aniket Ghorpade, Hansheng Zhu, Justin Qiu, Dea Rrozhani, Monica Lama, Mick Yang, Zixuan Bian, Ruohan Ren, Alan B. Hong, Jiatao Gu, Chris Callison-Burch

机构 * University of Pennsylvania（宾夕法尼亚大学） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.CL

State of Health Estimation of Batteries Using a Time-Informed Dynamic Sequence-Inverted Transformer

Janak M. Patel, Milad Ramezankhani, Anirudh Deodhar, Dagnachew Birru

机构 * Applied Research, Quantiphi（Quantiphi应用研究） 纠错

专题命中 其他安全 ：safety(abstract)；分类 cs.LG

Comments 11 pages, 3 figures

Human-Centered AI and Autonomy in Robotics: Insights from a Bibliometric Study

Simona Casini, Pietro Ducange, Francesco Marcelloni, Lorenzo Pollini

机构 * Department of Information Engineering, University of Pisa（信息工程系，比萨大学） 纠错

专题命中 其他安全 ：safety(abstract)；分类 cs.AI

Comments International Joint Conference on Neural Network 2025 - Accepted

Journal ref 10.1109/IJCNN64981.2025

Rapid Machine Learning-Driven Detection of Pesticides and Dyes Using Raman Spectroscopy

Quach Thi Thai Binh, Thuan Phuoc, Xuan Hai, Thang Bach Phan, Vu Thi Hanh Thu, Nguyen Tuan Hung

机构 * Faculty of Physics and Physics Engineering, University of Science, Ho Chi Minh City 700000, Viet Nam（物理系和物理工程系，科学大学，胡志明市700000，越南） ； Center for Innovative Materials and Architectures (INOMAR)（创新材料与架构中心） ； Department of Materials Science and Engineering, National Taiwan University, Taipei 10617, Taiwan（材料科学与工程系，台湾国立大学，台北10617，台湾） 纠错

专题命中 其他安全 ：safety(abstract)；分类 cs.LG

Comments 25 pages, 9 figures

Flash-Fusion: Enabling Expressive, Low-Latency Queries on IoT Sensor Streams with LLMs

Kausar Patherya, Ashutosh Dhekne, Francisco Romero

机构 * Georgia Institute of Technology（佐治亚理工学院） 纠错

专题命中 其他安全 ：safety(abstract)；分类 cs.AI

Comments 12 pages, 5 figures. Under review

Image-POSER: Reflective RL for Multi-Expert Image Generation and Editing

Hossein Mohebbi, Mohammed Abdulrahman, Yanting Miao, Pascal Poupart, Suraj Kothawade

机构 * University of Waterloo（滑铁卢大学） ； Vector Institute（向量研究所） ； Google（谷歌） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

LangGPS: Language Separability Guided Data Pre-Selection for Joint Multilingual Instruction Tuning

Yangfan Ye, Xiaocheng Feng, Xiachong Feng, Lei Huang, Weitao Ma, Qichen Hong, Yunfei Lu, Duyu Tang, Dandan Tu, Bing Qin

专题命中 其他安全 ：alignment(abstract)；分类 cs.CL

Comments AAAI2026 Main Track Accepted

Anomagic: Crossmodal Prompt-driven Zero-shot Anomaly Generation

Yuxin Jiang, Wei Luo, Hui Zhang, Qiyu Chen, Haiming Yao, Weiming Shen, Yunkang Cao

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

Language Drift in Multilingual Retrieval-Augmented Generation: Characterization and Decoding-Time Mitigation

Bo Li, Zhenghua Xu, Rui Xie

专题命中 其他安全 ：alignment(abstract)；分类 cs.CL

Comments AAAI'26, Oral Paper

Difference Vector Equalization for Robust Fine-tuning of Vision-Language Models

Satoshi Suzuki, Shin'ya Yamaguchi, Shoichiro Takeda, Taiga Yamane, Naoki Makishima, Naotaka Kawata, Mana Ihori, Tomohiro Tanaka, Shota Orihashi, Ryo Masumura

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

Comments Accepted by AAAI 2026

Remodeling Semantic Relationships in Vision-Language Fine-Tuning

Xiangyang Wu, Liu Liu, Baosheng Yu, Jiayan Qiu, Zhenwei Shi

机构 * Hangzhou International Innovation Institute, Beihang University（北京航空航天大学杭州国际创新研究院） ； School of Artificial Intelligence, Beihang University（北京航空航天大学人工智能学院） ； Nanyang Technological University（南洋理工大学） ； University of Leicester（莱斯特大学） ； School of Astronautics, Beihang University（北京航空航天大学航天学院） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.AI

Beyond Perplexity: Let the Reader Select Retrieval Summaries via Spectrum Projection Score

Zhanghao Hu, Qinglin Zhu, Siya Qi, Yulan He, Hanqi Yan, Lin Gui

专题命中 其他安全 ：alignment(abstract)；分类 cs.CL

Comments Accepted by AAAI 2026 Oral. Project link: https://zhanghao-aaai2026-sps.github.io/AAAI2026-SPS/

Generalized Linear Mode Connectivity for Transformers

Alexander Theus, Alessandro Cabodi, Sotiris Anagnostidis, Antonio Orvieto, Sidak Pal Singh, Valentina Boeva

专题命中 其他安全 ：alignment(abstract)；分类 cs.LG

Comments Accepted to NeurIPS 2025 (Oral)

Steering Pretrained Drafters during Speculative Decoding

Frédéric Berdoz, Peer Rheinboldt, Roger Wattenhofer

专题命中 其他安全 ：alignment(abstract)；分类 cs.LG

Comments Accepted at AAAI 2026

Contextual morphologically-guided tokenization for Latin encoder models

Marisa Hudspeth, Patrick J. Burns, Brendan O'Connor

机构 * Manning College of Information & Computer Sciences, University of Massachusetts Amherst（马萨诸塞大学阿默斯特分校信息与计算机科学学院） ； Institute for the Study of the Ancient World, New York University（纽约大学古代世界研究所） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.CL

Re-envisioning Euclid Galaxy Morphology: Identifying and Interpreting Features with Sparse Autoencoders

John F. Wu, Michael Walmsley

机构 * Space Telescope Science Institute（太空望远镜科学研究所） ； Johns Hopkins University（约翰霍普金斯大学） ； University of Toronto（多伦多大学） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.LG

Comments Authors contributed equally to this work. Accepted to NeurIPS Machine Learning and the Physical Sciences Workshop. See trained model at https://huggingface.co/mwalmsley/euclid-rr2-mae, HuggingFace demo at https://huggingface.co/spaces/mwalmsley/euclid_masked_autoencoder, and code at https://github.com/jwuphysics/euclid-galaxy-morphology-saes

The Non-Linear Representation Dilemma: Is Causal Abstraction Enough for Mechanistic Interpretability?

Denis Sutter, Julian Minder, Thomas Hofmann, Tiago Pimentel

专题命中 其他安全 ：alignment(abstract)；分类 cs.LG

Comments NeurIPS 2025 (Spotlight)

Knowledge Graph Analysis of Legal Understanding and Violations in LLMs

Abha Jha, Abel Salinas, Fred Morstatter

专题命中 其他安全 ：safety(abstract)；分类 cs.CL

Misaligned by Design: Incentive Failures in Machine Learning

David Autor, Andrew Caplin, Daniel Martin, Philip Marx

机构 * Massachusetts Institute of Technology, Google Technology and Society Fellows program, and NBER（麻省理工学院、谷歌技术与社会 fellows 程序及国家经济研究局） ； New York University and NBER（纽约大学及国家经济研究局） ； University of California, Santa Barbara（加州大学圣塔芭芭拉分校） ； Louisiana State University（路易斯安那州立大学） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.LG

Large Language Models Do Multi-Label Classification Differently

Marcus Ma, Georgios Chochlakis, Niyantha Maruthu Pandiyan, Jesse Thomason, Shrikanth Narayanan

机构 * University of Southern California（南加州大学） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.CL

Comments To be published in the Main Conference Proceedings of EMNLP 2025, 24 pages, 16 figures, 7 tables

ElastoGen: 4D Generative Elastodynamics

Yutao Feng, Yintong Shang, Xiang Feng, Lei Lan, Shandian Zhe, Tianjia Shao, Hongzhi Wu, Kun Zhou, Chenfanfu Jiang, Yin Yang

专题命中 其他安全 ：alignment(abstract)；分类 cs.LG

Self-Interpretability: LLMs Can Describe Complex Internal Processes that Drive Their Decisions

Dillon Plunkett, Adam Morris, Keerthi Reddy, Jorge Morales

机构 * Northeastern University（东北大学） ； Princeton University（普林斯顿大学） ； Independent Researcher（独立研究者） 纠错

专题命中 其他安全 ：safety(abstract)；分类 cs.CL

Quality Over Quantity? LLM-Based Curation for a Data-Efficient Audio-Video Foundation Model

Ali Vosoughi, Dimitra Emmanouilidou, Hannes Gamper

机构 * University of Rochester（罗切斯特大学） ； Microsoft Research（微软研究院） 纠错

专题命中 其他安全 ：alignment(abstract)；分类 cs.CL

Comments Accepted at EUSIPCO 2025 - 5 pages, 5 figures, 2 tables

Leveraging Text-Driven Semantic Variation for Robust OOD Segmentation

Seungheon Song, Jaekoo Lee

专题命中 其他安全 ：safety(abstract)；分类 cs.AI

Comments 8 pages, 5 figure references, 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) submission