AAAI

Your AI-Generated Image Detector Can Secretly Achieve SOTA Accuracy, If Calibrated

你的AI生成图像检测器可以秘密实现SOTA精度，如果校准

Muli Yang, Gabriel James Goenawan, Henan Wang, Huaiyuan Qin, Chenghao Xu, Yanhua Yang, Fen Fang, Ying Sun, Joo-Hwee Lim, Hongyuan Zhu

AI总结 本文提出基于贝叶斯决策理论的后处理校准框架，通过可学习标量修正提升AI生成图像检测的鲁棒性，无需重新训练。

Comments AAAI 2026. Code: https://github.com/muliyangm/AIGI-Det-Calib

Unmediated AI-Assisted Scholarly Citations

无中介的人工智能辅助学术引用

Stefan Szeider

AI总结 本文提出一种结合语言模型与直接数据库访问的架构，通过模型上下文协议实现准确的学术引用处理，提升文献检索的可靠性和效率。

Journal ref Open Conference Proceedings, Vol. 8 (2026): The Second Bridge on Artificial Intelligence for Scholarly Communication (AAAI-26)

A Lightweight Sparse Interaction Network for Time Series Forecasting

一种轻量级稀疏交互网络用于时间序列预测

Xu Zhang, Qitong Wang, Peng Wang, Wei Wang

AI总结 本文提出了一种轻量级稀疏交互网络LSINet，通过稀疏性诱导的伯努利分布学习时间步之间的关键连接，以提升时间序列预测的准确性和效率。

Comments The paper is published in AAAI Conference on Artificial Intelligence, AAAI 2025. The code is available at the link https://github.com/Meteor-Stars/LSINet

SAPO: Self-Adaptive Process Optimization Makes Small Reasoners Stronger

SAPO：自适应过程优化使小型推理器更强

Kaiyuan Chen, Guangmin Zheng, Jin Wang, Xiaobing Zhou, Xuejie Zhang

AI总结 SAPO通过自适应过程优化方法，有效缩小推理器与验证器之间的差距，提升小型语言模型在数学和代码任务中的性能。

Comments Accepted by AAAI 2026

CrossCheck-Bench: Diagnosing Compositional Failures in Multimodal Conflict Resolution

CrossCheck-Bench：多模态冲突解决中的组成性故障诊断

Baoliang Tian, Yuxuan Si, Jilong Wang, Lingyao Li, Zhongyuan Bao, Zineng Zhou, Tao Wang, Sixu Li, Ziyao Xu, Mingze Wang, Zhouzhuo Zhang, Zhihao Wang, Yike Yun, Ke Tian, Ning Yang, Minghui Qiu

AI总结 CrossCheck-Bench通过多阶段注释流程构建的基准测试，揭示了多模态模型在处理跨模态冲突时的瓶颈，并表明融合符号推理与视觉处理的方法能提升模型的稳健性。

Comments Accepted by AAAI 2026

Personalized Federated Learning with Bidirectional Communication Compression via One-Bit Random Sketching

通过一比特随机草图实现双向通信压缩的个性化联邦学习

Jiacheng Cheng, Xu Zhang, Guanghui Qiu, Yifang Zhang, Yinchuan Li, Kaiyuan Feng

AI总结 本文提出pFed1BS框架，通过一比特随机草图实现个性化联邦学习中的通信压缩，结合符号正则化和快速哈达玛变换，有效降低通信成本并保持性能

Comments Accepted in AAAI 2026

DeepGB-TB: A Risk-Balanced Cross-Attention Gradient-Boosted Convolutional Network for Rapid, Interpretable Tuberculosis Screening

DeepGB-TB: 一种风险平衡的跨注意力梯度提升卷积网络用于快速、可解释的肺结核筛查

Zhixiang Lu, Yulong Li, Feilong Tang, Zhengyong Jiang, Chong Li, Mian Zhou, Tenglong Li, Jionglong Su

AI总结 DeepGB-TB通过结合跨注意力机制和梯度提升决策树，实现快速、可解释的肺结核筛查，具有高准确率和低资源需求。

Comments Accepted by AAAI 2026 (oral)

Journal ref Proceedings of the AAAI Conference on Artificial Intelligence, 2026

KVmix: Gradient-Based Layer Importance-Aware Mixed-Precision Quantization for KV Cache

KVmix: 基于梯度的层重要性感知的KV缓存混合精度量化

Fei Li, Song Liu, Weiguo Wu, Shiqiang Nie, Jinyu Wang

机构 * Fei Li, Song Liu, Weiguo Wu, Shiqiang Nie, Jinyu Wang（作者） 纠错

AI总结 KVmix通过基于梯度的重要性分析实现KV缓存的混合精度量化，有效平衡精度与效率，实现低内存高吞吐的LLM推理

Comments AAAI 2026 Oral

Minimizing Inequity in Facility Location Games

在设施选址游戏中最小化不平等

Yuhang Guo, Houyu Zhou

AI总结 本文提出两种新的机制以最小化设施选址游戏中群体效应，统一了多个经典诚实机制，并改进了群体公平目标的近似界限。

Comments Accepted in AAAI 2026

CoRe-Fed: Bridging Collaborative and Representation Fairness via Federated Embedding Distillation

CoRe-Fed: 通过联邦嵌入蒸馏实现协同与表示公平性

Noorain Mukhtiar, Adnan Mahmood, Quan Z. Sheng

AI总结 CoRe-Fed通过联邦嵌入蒸馏实现协同与表示公平性，通过嵌入级正则化和公平性感知聚合策略提升模型性能和公平性。

Comments 7 pages (main content), 2 pages (references), Accepted in AAAI 2026

Kernelized Edge Attention: Addressing Semantic Attention Blurring in Temporal Graph Neural Networks

核化边注意力：解决时间图神经网络中语义注意力模糊问题

Govind Waghmare, Srini Rohan Gujulla Leel, Nikhil Tumbde, Sumedh B G, Sonia Gupta, Srikanta Bedathur

AI总结 KEAT通过核化边注意力机制提升时间图神经网络对时间依赖性的捕捉能力，实现更准确和可解释的信息传递。

Comments Accepted at AAAI 2026

Reinforced Rate Control for Neural Video Compression via Inter-Frame Rate-Distortion Awareness

通过帧间速率-失真意识的强化率控实现神经视频压缩

Wuyang Cong, Junqi Shi, Lizhong Wang, Weijing Shi, Ming Lu, Hao Chen, Zhan Ma

AI总结 本文提出基于强化学习的神经视频压缩率控方法，通过联合优化比特率分配和编码参数，实现更高效的压缩性能和更低的开销。

Comments Accepted by AAAI 2026

PSM: Prompt Sensitivity Minimization via LLM-Guided Black-Box Optimization

PSM：通过LLM引导的黑盒优化实现提示敏感性最小化

Huseein Jawad, Nicolas Brunel

AI总结 本文提出PSM框架，通过LLM引导的黑盒优化，利用SHIELD层减少系统提示泄露，提升LLM安全性和隐私保护。

Comments accepted at the Trustworthy Agentic AI Workshop @ AAAI 2026 (Singapore, Jan 27, 2026). Workshop PDF: https://trustagenticai.github.io/AAAI2026/AAAI-Workshop/20.pdf

Efficient Thought Space Exploration Through Strategic Intervention

通过战略干预实现高效思维空间探索

Ziheng Li, Hengyi Cai, Xiaochi Wei, Yuchen Li, Shuaiqiang Wang, Zhi-Hong Deng, Dawei Yin

机构 * Baidu（百度） 纠错

AI总结 HPR框架通过战略干预减少计算成本，在保持高准确率的同时实现高效推理。

Comments AAAI 2026

Beyond Cosine Similarity: Magnitude-Aware CLIP for No-Reference Image Quality Assessment

超越余弦相似度：面向无参考图像质量评估的幅度感知CLIP

Zhicheng Liao, Dongxu Wu, Zhenshan Shi, Sijie Mai, Hanwei Zhu, Lingyu Zhu, Yuncheng Jiang, Baoliang Chen

AI总结 本文提出一种基于CLIP的无参考图像质量评估方法，通过引入幅度感知的特征处理和置信度引导的融合机制，提升图像质量评估的准确性。

Comments Accepted by AAAI 2026

Enhancing Human-Like Responses in Large Language Models

增强大型语言模型的人类样响应

Ethem Yağız Çalık, Talha Rüzgar Akkuş

AI总结 本文提出通过提升自然语言理解、对话连贯性和情感智能来增强大型语言模型的人类化响应，展示了改进用户交互和跨领域应用的潜力。

Comments Presented at the AAAI-26 Workshop on Personalization in the Era of Large Foundation Models (PerFM), Singapore, January 2026

Journal ref Presented at the AAAI-26 Workshop on Personalization in the Era of Large Foundation Models (PerFM), 2026