AI资讯日报 - 2026/8/26

👨‍🔬 Andreas Hochlehnert, Marianna Nezhurina, Mehdi Cherti, Andrej Radonjic, Thaddäus Wiedemer, Christoph Schuhmann, Romain Beaumont, Wieland Brendel, Bernhard Schölkopf, A. Sophia Koepke, Jenia Jitsev, Matthias Bethge

CAFE: Self-Improving Search Agents Need Co-Evolving Feedback

学术论文 ArXiv 重要度: 9

提出CAFE框架，代理与评论者共同进化，通过在线和离线优化提升搜索代理性能，减少幻觉。

👨‍🔬 Boyang Liu, Senjie Jin, Peixin Wang, Zhangyue Yin, Yibo Wang, Yuhao Zhou, Xinbing Liang, Shizheng Zhu, Yuhui Wang, Jingqi Tong, Zhiheng Xi, Jiazheng Zhang, Clive Bai, Clarenceai, Blaze Chen, Tao Gui, Qi Zhang, Xuanjing Huang

Right Diagnoses, Decorative Reasoning:A Perturbation Audit of Medical Chain-of-Thought

学术论文 ArXiv 重要度: 9

通过扰动审计发现医疗链式思维推理与答案脱节，提出CDR指标量化，多数模型存在不忠实推理。

👨‍🔬 Mengzhu Xu, Jifan Gao, Xia Jiang, Yaoxin Wu, Xi Long

FedV-KGQA: Multi-Hop Question Answering over Vertically Partitioned Knowledge Graphs

学术论文 ArXiv 重要度: 8

提出FedV-KGQA框架，在垂直划分的知识图谱上实现多跳问答，无需跨机构通信，性能接近集中式方法。

👨‍🔬 Md Saikat Islam Khan Bappy, Oshani Seneviratne

Reading Is Not Using: Retrieval, Judgment, and the Design of AI Financial Research Workflows

学术论文 ArXiv 重要度: 8

揭示金融分析中LLM检索与判断脱节的问题，提出工作流设计影响信息使用，结构化重述可恢复风险披露影响。

👨‍🔬 Miao Liu, Zhizhe Liu

StarHarness: Evolving Harnesses with Stratified Search for Enterprise Environments

学术论文 ArXiv 重要度: 8

提出StarHarness，通过分层搜索进化环境特定代理框架，显著提升企业任务性能，且可跨模型迁移。

👨‍🔬 Esakkivel Esakkiraja, Denis Akhiyarov, Vikas Yadav, Sai Rajeswar, Patrice Bechard, Sridhar Nemala, Sagar Davasam

A Dual-Dimensional LLM Framework for Automated Item Incidental Content Similarity Analysis in Large-Scale Assessments

学术论文 ArXiv 重要度: 7

提出双维度LLM框架分析试题非故意内容相似性，改善自适应测试中的项目选择与估计稳定性。

👨‍🔬 Jing Huang, Jihong Zhang, Hua-Hua Chang

Constrained Entity Selection under Partial Knowledge for LLM-Based Knowledge Graph QA

学术论文 ArXiv 重要度: 7

提出CES-PK问题，利用符号约束过滤无效答案，提高KGQA精度并保持召回率，适用于不完整知识图谱。

👨‍🔬 Emanuel Kitzelmann

Strictly Causal Streaming Video Anomaly Detection with a Theoretically-Grounded State-Space Core

学术论文 ArXiv 重要度: 7

提出严格因果流式视频异常检测器，基于状态空间模型，实时处理且延迟低，但在准确率上仍有提升空间。

👨‍🔬 Yogesh Kumar

Automatic Model Card Generation Using an LLM

学术论文 ArXiv 重要度: 6

利用LLM自动生成和整理模型卡片，标准化结构，提高可比较性，生成质量依赖资源可用性。

👨‍🔬 Tajkia Rahman Toma, Balreet Grewal, Cor-Paul Bezemer

🤖 AI资讯日报

📊 今日趋势总结

The AI Crackpot Index

Ask HN: Is the rate of progress in AI exponential?

MIT Non-AI License

Ask HN: Anyone concerned about NYC Local Law 144?

Ask HN: What's the pain using current AI algorithms?

Why Boring Businesses Outlast AI Hype Cycles

Ask HN: What would you read to learn about "artificial intelligence"?

Common Lisp + Machine Learning Internship at Google (Mountain View, CA)

Show HN: Startup Raising capital through Book Sales

The Next Bill Gates or Albert Einstein in AI “Chris Clark” – Yourobot

Bioinformatician

NLP, AI, ML, bots – a passing trend or much more? What's your take on this?

Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses

SPO++: Stream-Aligned Policy Optimization for Asynchronous Agentic RL

LAION-BVD: A 10-Million-Hour Open Video Dataset for Multimodal Pre-training

CAFE: Self-Improving Search Agents Need Co-Evolving Feedback

Right Diagnoses, Decorative Reasoning:A Perturbation Audit of Medical Chain-of-Thought

FedV-KGQA: Multi-Hop Question Answering over Vertically Partitioned Knowledge Graphs

Reading Is Not Using: Retrieval, Judgment, and the Design of AI Financial Research Workflows

StarHarness: Evolving Harnesses with Stratified Search for Enterprise Environments

A Dual-Dimensional LLM Framework for Automated Item Incidental Content Similarity Analysis in Large-Scale Assessments

Constrained Entity Selection under Partial Knowledge for LLM-Based Knowledge Graph QA

Strictly Causal Streaming Video Anomaly Detection with a Theoretically-Grounded State-Space Core

Automatic Model Card Generation Using an LLM

📅 历史日报目录