AAAI

Conformal Coverage Guarantees for Any Video Temporal Grounder

任意视频时间定位器的共形覆盖保证

Aseel Mohamed, Rasul Khanbayov, Erchin Serpedin, Hasan Kurban

AI总结 针对视频时间定位器的模糊性问题，提出与模型无关的COVER包装器，可保证输出区域以至少1-α的概率包含真实时刻，在多基准和定位器上验证了其覆盖度符合目标值。

Comments Submitted to AAAI 2027

People Are Not Just Their Countries. Disentangling Social Determinants of LLM Value Alignment Across Europe

人不只是其所属国家：在欧洲范围内解耦大型语言模型（LLM）价值对齐的社会决定因素

Maria-Louisa Wightman, Guillaume Bied, Tijl De Bie

AI总结 本研究依托欧洲社会调查，发现LLMs与不同社会人口学群体的价值观对齐存在差异，国籍作为单独变量的解释力与全部社会人口学变量相当，国家与社会人口学因素在解释对齐模式上互补。

Comments Accepted at AIES 2026 (9th AAAI/ACM Conference on AI, Ethics, and Society)

Data Annotation as Measurement

数据标注即测量

Emma Harvey, Allison Koenecke, Rene F. Kizilcec

AI总结 本文将数据标注重新定义为测量，通过文献综述和访谈开发标注问题诊断框架，明确标注问题来源并给出超越一致性的质量评估方法，为提升AI标注数据质量提供概念基础。

Comments AAAI/ACM Conference on AI, Ethics, and Society (AIES 2026)

Calibrating WEAT Against Anisotropy: ZCA Whitening as a Geometric Pre-Processing Step for Embedding Association Tests

针对各向异性校准WEAT：将ZCA白化作为嵌入关联测试的几何预处理步骤

Seitaro Ono, Senna Ross, Jun Saiki

AI总结 本研究提出将ZCA白化作为预处理步骤校准WEAT，可降低嵌入空间各向异性，超30%WEAT结果显著性改变，提升语义相似度，为相关偏差测量提供更可靠基础。

Comments Extended version (with appendices) of a paper accepted at the 9th AAAI/ACM Conference on AI, Ethics, and Society (AIES 2026)

On Seeding Watermarks to Detect Verbatim LLM Copy-Paste Responses

检测作业中的直接LLM复制粘贴

Aizierjiang Aiersilan, Artin Yousefi, Robert Pless

机构 * The George Washington University（乔治·华盛顿大学） 纠错

AI总结 本文提出SteganoPrompt工具，通过在作业提示中嵌入隐形指令，使LLM在响应时生成特征签名，帮助教师检测学生直接复制粘贴模型回复的行为。

Comments We are pleased to announce that this paper has been accepted by the 9th AAAI/ACM Conference on AI, Ethics, and Society (AIES 2026). We appreciate the valuable feedback from the reviewers and look forward to sharing our findings with the community

Judge a Book by its Cover: Investigating Multi-Modal LLMs for Multi-Page Handwritten Document Transcription

通过封面判断书籍：调查多模态大语言模型用于多页手写文档转录

Benjamin Gutteridge, Matthew Thomas Jackson, Toni Kukurin, Xiaowen Dong

机构 * University of Oxford（牛津大学） ； QuantCo 纠错

AI总结 本文研究多模态大语言模型在多页手写文档转录中的应用，提出OCR+PAGE-1和OCR+PAGE-N策略，通过共享页面内容提升转录效果。

Comments 10 pages (36 including references and appendices), 11 figures, accepted at COLM 2026, earlier version accepted at AAAI 2025 Workshop on Document Understanding and Intelligence

Mitigating Negative Flips via Margin Preserving Training

Simone Ricci, Niccolò Biondi, Federico Pernici, Alberto Del Bimbo

Comments Accepted at AAAI2026

Journal ref Proceedings of the AAAI Conference on Artificial Intelligence, 40(11), 8721-8730, 2026

An Empirical Analysis of Constrained Support Vector Quantile Regression for Nonparametric Probabilistic Forecasting of Wind Power

风电非参数概率预测的带约束支持向量分位数回归的实证分析

Kostas Hatalis, Shalinee Kishore, Katya Scheinberg, Alberto Lamadrid

AI总结 本文结合支持向量机、非线性分位数回归与非交叉约束提出风电非参数概率预测方法，基于全球能源预测竞赛2014年数据验证，该方法性能优于三个基准模型且避免分位数估计重叠问题。

Comments Originally published at The AAAI-17 Workshop on Artificial Intelligence for Smart Grids and Smart Buildings

Journal ref Thirty-First AAAI Conference on Artificial Intelligence, 2017