arXivDaily arXiv每日学术速递 周一至周五更新

期刊&会议

NeurIPS

Conference on Neural Information Processing Systems · 会议 · Machine Learning

2026-06-25 至 2026-06-25 共收录 7
2606.24965 2026-06-25 cs.AI cs.LG 新提交

Project Auto-World: Towards Automated Benchmarking of Neural Relational Reasoners

Project Auto-World: 迈向神经关系推理器的自动化基准测试

Anirban Das, Joanne Boisson, Irtaza Khalid, Sumita Garai, Steven Schockaert

机构 * Cardiff University(卡迪夫大学) University of Pennsylvania(宾夕法尼亚大学)

AI总结 利用大语言模型自动化生成基准测试实例,通过进化搜索和自主代理搜索发现困难样本,提升Edge Transformer的泛化能力,并应用于新世界以推动神经关系推理的自主研究。

Comments Submitted to NeurIPS 2026 E&D track. Code is available at https://github.com/autoworldrules/auto-world-rules

详情

展开后加载摘要…

URL PDF HTML 收藏
2606.25042 2026-06-25 cs.IT math.IT math.PR math.ST stat.ML stat.TH 新提交

Information from coincidences

来自重合的信息

Akshay Balsubramani

AI总结 本文证明了一个统一的代数混合重合恒等式,将信息论中多个变分结果(如Sanov分解、Chernoff信息、Donsker-Varadhan不等式等)作为特例,并推广到多先验情形,应用于语言模型和基因组序列分析。

Comments 78 pages, 16 figures, 7 tables. Submitted to NeurIPS 2026. A mixed coincidence partition function gives Sanov, Chernoff, PAC-Bayes, and Renyi as corollaries

详情

展开后加载摘要…

URL PDF HTML 收藏
2606.24187 2026-06-25 cs.CV 新提交

Towards Fast and Effective Long Video Understanding of Multimodal Large Language Models via Adaptive Quasi-Gaussian Sampling

面向多模态大语言模型的长视频快速有效理解:自适应准高斯采样

Kun Zhang, Chenxin Fang, Tao Chen, Baiyang Song, Yunhang Shen, Yiyi Zhou, Rongrong Ji

机构 * Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, Xiamen University(厦门大学多媒体可信感知与高效计算教育部重点实验室)

AI总结 提出自适应无训练帧采样方法AdaQ,基于高斯分布3-σ规则动态调整采样区间,在仅用64帧下使Qwen3-VL-8B平均超越GPT4o 15.8%,显著提升长视频理解的鲁棒性和效率。

Comments NeurIPS 2026 submission. 15 pages, 8 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.17037 2026-06-25 cs.CV cs.AI 版本更新

AMVICC: A Novel Benchmark for Cross-Modal Failure Mode Profiling for VLMs and IGMs

AMVICC: 一种用于VLM和IGM跨模态故障模式分析的新型基准

Aahana Basappa, Pranay Goel, Anusri Karra, Anish Karra, Asa Gilmore, Kevin Zhu

机构 * Centennial High School, Frisco, Texas, USA(Centennial High School, Texas, USA) Lebanon Trail High School, Frisco, Texas, USA(Lebanon Trail High School, Texas, USA) West Windsor-Plainsboro High School, Princeton Junction, New Jersey, USA(West Windsor-Plainsboro High School, New Jersey, USA) Algoverse AI Research, Palo Alto, California, USA(Algoververse AI Research, California, USA)

AI总结 提出AMVICC基准,通过图像到文本和文本到图像任务系统比较多模态大模型和图像生成模型的视觉推理失败模式,发现故障模式在模型和模态间共享,但存在特定于模型和模态的失败。

Comments 14 pages, 4 figures, 8 tables. Presented at the 39th Conference on Neural Information Processing Systems Workshop: VLM4RWD. Presented at the 43th International Conference on Machine Learning Workshops: ICML 2026 CTB, ICML 2026 FAGEN, ICML 2026 EMM-QA. Authors Aahana Basappa and Pranay Goel contributed equally. Code: https://github.com/AahanaB24/AMVICC, Data: https://doi.org/10.5281/zenodo.17646068

详情

展开后加载摘要…

URL PDF HTML 收藏
2505.13731 2026-06-25 cs.CV 版本更新

GeoRanker: Distance-Aware Ranking for Worldwide Image Geolocalization

GeoRanker:面向全球图像地理定位的距离感知排序

Pengyue Jia, Seongheon Park, Song Gao, Xiangyu Zhao, Sharon Li

机构 * Department of Data Science, City University of Hong Kong(城市大学数据科学系) Department of Computer Sciences, University of Wisconsin-Madison(威斯康星大学麦迪逊分校计算机科学系) Department of Geography, University of Wisconsin-Madison(威斯康星大学麦迪逊分校地理系)

AI总结 提出GeoRanker框架,利用大视觉语言模型联合编码查询-候选交互并预测地理邻近性,引入多阶距离损失以建模结构化空间关系,在IM2GPS3K和YFCC4K基准上达到最优。

Comments NeurIPS 2025

详情

展开后加载摘要…

URL PDF HTML 收藏
2509.00704 2026-06-25 cs.LG cs.AI q-bio.QM 版本更新

Why Pool When You Can Flow? Active Learning with GFlowNets

为何池选,不如流选?基于GFlowNets的主动学习

Renfei Zhang, Mohit Pandey, Artem Cherkasov, Martin Ester

机构 * School of Computer Science, Simon Fraser University, Burnaby, BC, Canada(Simon Fraser大学计算机科学学院,Burnaby, BC, Canada) Vancouver Prostate Centre, University of British Columbia, Vancouver, BC, Canada(温哥华前列腺中心,不列颠哥伦比亚大学,Vancouver, BC, Canada) Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada(不列颠哥伦比亚大学医学院,Vancouver, BC, Canada) Diagen AI

AI总结 提出BALD-GFlowNet框架,用生成流网络直接采样高信息分子,替代传统池选,实现与池大小无关的可扩展性,在虚拟筛选中达到与BALD相当的性能并生成更多样化分子。

Comments Accepted at the NeurIPS 2025 Workshop on AI Virtual Cells and Instruments: A New Era in Drug Discovery and Development (AI4D3 2025), San Diego, California, USA. 6 pages; 5 figures

详情

展开后加载摘要…

URL PDF HTML 收藏
2506.05252 2026-06-25 cs.LG cs.GT cs.MA

Conservative classifiers do consistently well with improving agents: characterizing statistical and online learning

保守分类器在改进的智能体中表现一致:刻画统计和在线学习

Dravyansh Sharma, Alec Sun

机构 * Northwestern University(西北大学) Toyota Technological Institute at Chicago(芝加哥丰田技术研究所) Alphabetical order(字母顺序) University of Chicago(芝加哥大学)

AI总结 本文研究了改进智能体对分类器的影响,提出了一种非对称的最小一致概念类,并在可实现设置中精确刻画了带改进的proper学习。通过欧几里得球改进集,解决了开放性问题,降低了泛化误差并改进了在线学习界限。

Comments 26 pages

Journal ref Advances in Neural Information Processing Systems (2025)

详情

展开后加载摘要…

URL PDF HTML 收藏