AAAI

Faster Certified Symmetry Breaking Using Orders With Auxiliary Variables

更快的利用辅助变量的顺序进行认证对称性打破

Markus Anders, Bart Bogaerts, Benjamin Bogø, Arthur Gontier, Wietze Koops, Ciaran McCreesh, Magnus O. Myreen, Jakob Nordström, Andy Oertel, Adrian Rebola-Pardo, Yong Kiam Tan

AI总结 本文提出了一种利用辅助变量编码顺序的方法，以提高对称性打破的认证效率。

Comments 26 pages. Extended version (with appendix) of the paper to appear in AAAI 2026

D-GARA: A Dynamic Benchmarking Framework for GUI Agent Robustness in Real-World Anomalies

D-GARA：一种用于真实世界异常中GUI代理鲁棒性评估的动态基准框架

Sen Chen, Tong Zhao, Yi Bin, Fei Ma, Wenqi Shao, Zheng Wang

AI总结 D-GARA提出一种动态基准框架，用于评估Android GUI代理在真实世界异常中的鲁棒性，揭示了现有代理在异常环境下的性能下降问题。

Comments Accepted to AAAI 2026

Formal Abductive Latent Explanations for Prototype-Based Networks

基于原型网络的正式归纳性隐式解释

Jules Soria, Zakaria Chihani, Julien Girard-Satabin, Alban Grastien, Romain Xu-Darme, Daniela Cancila

AI总结 本文提出基于原型网络的正式归纳性隐式解释方法，通过形式化方法提升模型可解释性，适用于图像分类任务。

Comments Accepted at AAAI-26

ECPv2: Fast, Efficient, and Scalable Global Optimization of Lipschitz Functions

ECPv2: 快速、高效且可扩展的Lipschitz函数全局优化

Fares Fourati, Mohamed-Slim Alouini, Vaneet Aggarwal

AI总结 ECPv2通过引入自适应下界、Worst-m记忆机制和固定随机投影，实现高维非凸函数优化的高效且可扩展的全局优化方法。

Comments Accepted at AAAI 2026 (main technical track), extended version

NutriScreener: Retrieval-Augmented Multi-Pose Graph Attention Network for Malnourishment Screening

NutriScreener: 一种结合检索增强的多姿态图注意力网络用于营养不良筛查

Misaal Khan, Mayank Vatsa, Kuldeep Singh, Richa Singh

AI总结 NutriScreener通过结合检索增强的多姿态图注意力网络，实现儿童营养不良的高效检测与体格测量预测，适用于低资源环境。

Comments Accepted in AAAI 2026 Special Track on AI for Social Impact

Large Language Model-Based Reward Design for Deep Reinforcement Learning-Driven Autonomous Cyber Defense

基于大型语言模型的深度强化学习驱动自主网络防御的奖励设计

Sayak Mukherjee, Samrat Chatterjee, Emilie Purvine, Ted Fujimoto, Tegan Emerson

AI总结 本文提出利用大型语言模型设计奖励，以提升深度强化学习在自主网络防御中的效果。

Comments Accepted in the AAAI-26 Workshop on Artificial Intelligence for Cyber Security (AICS)

LLaVA$^3$: Representing 3D Scenes like a Cubist Painter to Boost 3D Scene Understanding of VLMs

LLaVA$^3$：像立体主义画家一样表示3D场景以提升VLM的3D场景理解

Doriand Petit, Steve Bourgeois, Vincent Gay-Bellile, Florian Chabot, Loïc Barthe

AI总结 LLaVA$^3$通过立体主义方法提升VLM对3D场景的理解能力，利用多视角2D图像无需微调实现更优的3D场景理解。

Comments Accepted at AAAI'26

TOFA: Training-Free One-Shot Federated Adaptation for Vision-Language Models

TOFA: 无需训练的一次性联邦适应用于视觉-语言模型

Li Zhang, Zhongxuan Han, XiaoHua Feng, Jiaming Zhang, Yuyuan Li, Linbo Jiang, Jianan Lin, Chaochao Chen

AI总结 TOFA提出一种无需训练的一次性联邦适应方法，通过视觉和文本管道提取任务相关表示，以高效适应视觉-语言模型在联邦学习中的应用。

Comments Accepted by AAAI 2026

Incorporating Self-Rewriting into Large Language Model Reasoning Reinforcement

将自我修正纳入大语言模型推理强化

Jiashu Yao, Heyan Huang, Shuang Zeng, Chuwei Luo, WangJie You, Jie Tang, Qingsong Liu, Yuhang Guo, Yangyang Kang

机构 * ByteDance China（字节跳动中国） 纠错

AI总结 本研究提出自我修正框架，通过重写推理文本提升大语言模型的内部推理质量，实验表明其在准确性与推理长度权衡上优于现有方法。

Comments Accepted to AAAI 2026

vMFCoOp: Towards Equilibrium on a Unified Hyperspherical Manifold for Prompting Biomedical VLMs

vMFCoOp：在统一超球面流形上实现平衡的提示生物医学VLMs

Minye Shao, Sihan Guo, Xinrun Li, Xingyu Miao, Haoran Duan, Yang Long

AI总结 vMFCoOp通过统一超球面流形上的vMF分布估计，实现LLM与CLIP主干间的语义对齐，提升生物医学VLMs的提示效果和少样本分类性能。

Comments Accepted as an Oral Presentation at AAAI 2026 Main Technical Track (this version is not peer-reviewed; it is the extended version)

AgentSwift: Efficient LLM Agent Design via Value-guided Hierarchical Search

AgentSwift: 通过价值引导的分层搜索高效设计LLM代理

Yu Li, Lehui Li, Zhihao Wu, Qingmin Liao, Jianye Hao, Kun Shao, Fengli Xu, Yong Li

AI总结 AgentSwift通过价值引导的分层搜索高效设计LLM代理，实现8.34%的性能提升。

Comments AAAI-2026

GazeInterpreter: Parsing Eye Gaze to Generate Eye-Body-Coordinated Narrations

GazeInterpreter: 解析眼动以生成眼-身体协调的叙述

Qing Chang, Zhiming Hu

AI总结 GazeInterpreter通过解析眼动数据生成眼-身体协调的叙述，提升人类行为理解的准确性与效率。

Comments Accepted to AAAI 2026. 9 pages, 4 figures

FT-NCFM: An Influence-Aware Data Distillation Framework for Efficient VLA Models

FT-NCFM: 一种面向影响的数据蒸馏框架用于高效VLA模型

Kewei Chen, Yayu Long, Shuai Li, Mingsheng Shang

AI总结 FT-NCFM通过智能数据蒸馏框架提升VLA模型效率，实现训练数据减少80%的同时成功率达85-90%

Comments Accepted at the AAAI Conference on Artificial Intelligence (AAAI-26)

PIPHEN: Physical Interaction Prediction with Hamiltonian Energy Networks

PIPHEN: 基于哈密顿能量网络的物理交互预测

Kewei Chen, Yayu Long, Mingsheng Shang

AI总结 PIPHEN通过物理交互预测网络和哈密顿能量网络，实现多机器人系统的高效协同控制，显著降低通信延迟并提高任务成功率。

Comments Accepted at the AAAI Conference on Artificial Intelligence (AAAI-26)

A Switching Framework for Online Interval Scheduling with Predictions

一种基于预测的在线区间调度切换框架

Antonios Antoniadis, Ali Shahheidar, Golnoosh Shahkarami, Abolfazl Soltani

机构 * University of Twente（特文特大学） ； Sharif University of Technology（谢赫大学） ； Max Planck Institut für Informatik, Universität des Saarlandes（马克斯·普朗克研究所，萨尔布吕克大学） 纠错

AI总结 本文提出SemiTrust-and-Switch框架，结合预测和经典算法，平衡一致性与鲁棒性，提升在线区间调度性能。

Comments This paper will appear in AAAI 2026

Target Refocusing via Attention Redistribution for Open-Vocabulary Semantic Segmentation: An Explainability Perspective

通过注意力重新分配的目标重聚焦实现开放词汇语义分割：可解释性视角

Jiahao Li, Yang Lu, Yachao Zhang, Yong Xie, Fangyong Wang, Yuan Xie, Yanyun Qu

AI总结 本文提出RF-CLIP方法，通过模拟人类注意力重聚焦行为，解决CLIP在开放词汇语义分割中的多模态对齐问题，提升密集预测性能。

Comments Accepted by AAAI 2026

Simba: Towards High-Fidelity and Geometrically-Consistent Point Cloud Completion via Transformation Diffusion

Simba: 通过变换扩散实现高保真的几何一致点云补全

Lirui Zhang, Zhengkai Zhao, Zhi Zuo, Pan Gao, Jie Qin

AI总结 Simba通过变换扩散方法，结合对称先验和扩散模型，实现高保真的点云补全，解决传统方法的过拟合和噪声敏感问题，取得最先进性能。

Comments Accepted for publication at the 40th AAAI Conference on Artificial Intelligence (AAAI-26)

Real-Time 3D Object Detection with Inference-Aligned Learning

基于推理对齐学习的实时3D物体检测

Chenyu Zhao, Xianwei Zheng, Zimin Xia, Linwei Yue, Nan Xue

AI总结 SR3D通过空间优先级最优传输分配和排名意识自适应自蒸馏方案，提升实时3D物体检测的准确性和效率

Comments Accepted by AAAI 2026

VTinker: Guided Flow Upsampling and Texture Mapping for High-Resolution Video Frame Interpolation

VTinker: 用于高分辨率视频帧插值的引导流上采样与纹理映射

Chenyang Wu, Jiayi Fu, Chun-Le Guo, Shuhao Han, Chongyi Li

AI总结 VTinker通过引导流上采样和纹理映射技术，提升高分辨率视频帧插值的精度与质量。

Comments Accepted by AAAI 2026

Multi-Faceted Attack: Exposing Cross-Model Vulnerabilities in Defense-Equipped Vision-Language Models

多面攻击：揭示配备防御机制的视觉语言模型中的跨模型漏洞

Yijun Yang, Lichao Wang, Jianping Zhang, Chi Harold Liu, Lanqing Hong, Qiang Xu

AI总结 多面攻击揭示了配备防御机制的视觉语言模型中的跨模型安全漏洞，通过注意力转移攻击和轻量级转移增强算法，实现了58.5%的成功率。

Comments AAAI 2026 Oral

Efficient Chromosome Parallelization for Precision Medicine Genomic Workflows

高效染色体并行化用于精准医疗基因组工作流

Daniel Mas Montserrat, Ray Verma, Míriam Barrabés, Francisco M. de la Vega, Carlos D. Bustamante, Alexander G. Ioannidis

AI总结 本文提出高效染色体并行化方法，通过动态调度器和静态调度器优化资源分配，减少内存溢出并提升基因组工作流执行效率。

Comments Accepted at AAAI 2026

InfoCLIP: Bridging Vision-Language Pretraining and Open-Vocabulary Semantic Segmentation via Information-Theoretic Alignment Transfer

InfoCLIP: 通过信息论对齐转移连接视觉语言预训练与开放词汇语义分割

Muyao Yuan, Yuanhong Zhang, Weizhan Zhang, Lan Ma, Yuan Gao, Jiangyong Ying, Yudeng Xin

AI总结 InfoCLIP通过信息论对齐转移提升开放词汇语义分割的性能，有效解决预训练CLIP在微调过程中的过拟合问题。

Comments Accepted by AAAI 2026

AquaSentinel: Next-Generation AI System Integrating Sensor Networks for Urban Underground Water Pipeline Anomaly Detection via Collaborative MoE-LLM Agent Architecture

AquaSentinel：下一代集成传感器网络的AI系统，通过协作MoE-LLM代理架构实现城市地下供水管道异常检测

Qiming Guo, Bishal Khatri, Wenbo Sun, Jinwen Tang, Hua Zhang, Wenlu Wang

AI总结 AquaSentinel通过协作MoE-LLM代理架构实现城市地下供水管道异常检测，结合物理信息和稀疏传感技术，以低成本实现高精度的实时监测。

Comments 7 pages, 1 figure, 2 tables, Accepted to the 40th AAAI Conference on Artificial Intelligence (AAAI 2026), IAAI Deployed Applications Track

TopoReformer: Mitigating Adversarial Attacks Using Topological Purification in OCR Models

TopoReformer：利用拓扑净化缓解OCR模型中的对抗攻击

Bhagyesh Kumar, A S Aravinthakashan, Akshat Satyanarayan, Ishaan Gakhar, Ujjwal Verma

AI总结 TopoReformer通过拓扑净化技术提升OCR模型对对抗攻击的鲁棒性，无需显式梯度正则化，有效缓解扰动并保持文本结构完整性。

Comments Accepted at AAAI 2026 AI for CyberSecurity (AICS) Workshop

Disagreement is Disappearing on U.S. Cable Debate Shows

分歧正在消失于美国有线辩论节目中

S M Mehedi Zaman, Kiran Garimella

AI总结 研究发现美国有线辩论节目中的分歧呈现下降趋势，揭示节目正从真实讨论转向党派确认，加剧社会极化。

Comments Accepted at the International AAAI Conference on Web and Social Media (ICWSM 2026)

Identifying the Supply Chain of AI for Trustworthiness and Risk Management in Critical Applications

识别用于可信性和风险管理的AI供应链在关键应用中的供应链

Raymond K. Sheh, Karen Geappen

AI总结 本文提出了一种分类法，用于分类AI供应链实体，以帮助利益相关者系统评估AI供应链风险，特别是在关键应用中

Comments Presented at the 2025 AAAI Fall Symposium - AI Trustworthiness and Risk Assessment for Challenged Contexts (ATRACC)

The Publication Choice Problem

出版选择问题

Haichuan Wang, Yifan Wu, Haifeng Xu

AI总结 本文提出出版选择问题模型，分析出版选择对影响力因子的双向影响，揭示聚光灯标注对不同出版物的影响差异。

Comments Accepted by AAAI 2026 as an oral presentation

Do Not Merge My Model! Safeguarding Open-Source LLMs Against Unauthorized Model Merging

不要合并我的模型！保护开源大语言模型免受未经授权的模型合并攻击

Qinfeng Li, Miao Pan, Jintao Chen, Fu Teng, Zhiqiang Shen, Ge Su, Hao Peng, Xuhong Zhang

AI总结 MergeBarrier通过破坏线性模式连接来主动防止未经授权的模型合并，实现高安全性和低性能损失的开源大语言模型保护

Comments Accepted by AAAI 2026 Conference

TubeRMC: Tube-conditioned Reconstruction with Mutual Constraints for Weakly-supervised Spatio-Temporal Video Grounding

TubeRMC: 基于互约束的管状重建用于弱监督空间-时间视频定位

Jinxuan Li, Yi Zhang, Jian-Fang Hu, Chaolei Tan, Tianming Liang, Beihao Xia

AI总结 TubeRMC通过引入互约束的管状重建方法，在弱监督条件下提升空间-时间视频定位的准确性和一致性。

Comments Accepted to AAAI 2026

MMVA: Multimodal Matching Based on Valence and Arousal across Images, Music, and Musical Captions

MMVA: 基于估值和唤醒度的多模态匹配

Suhwan Choi, Kyu Won Kim, Myungjoo Kang

AI总结 MMVA通过多模态匹配基于估值和唤醒度，实现跨图像、音乐和音乐字幕的情感内容捕捉，并在估值-唤醒度预测任务中取得最佳性能。

Comments Paper accepted in Artificial Intelligence for Music workshop at AAAI 2025