AI资讯日报 - 2026/5/3

👨‍🔬 Yujun Wu, Dongxu Zhang, Xinchen Li, Jinhang Xu, Yiling Duan, Yumou Liu, Jiabao Pan, Xuanhe Zhou, Jingxuan Wei, Siyuan Li, Jintao Chen, Conghui He, Cheng Tan

Claw-Eval-Live: A Live Agent Benchmark for Evolving Real-World Workflows

学术论文 ArXiv 重要度: 9

实时工作流智能体基准，基于动态需求构建任务，发现当前模型最高通过率仅66.7%。

👨‍🔬 Chenxin Li, Zhengyang Tang, Huangxin Lin, Yunlong Lin, Shijue Huang, Shengyuan Liu, Bowen Ye, Rang Li, Lei Li, Benyou Wang, Yixuan Yuan

Computing Equilibrium beyond Unilateral Deviation

学术论文 ArXiv 重要度: 8

提出最小化联盟偏离激励的均衡概念，保证存在性，并给出计算复杂度下界与匹配算法。

👨‍🔬 Mingyang Liu, Gabriele Farina, Asuman Ozdaglar

PhyCo: Learning Controllable Physical Priors for Generative Motion

学术论文 ArXiv 重要度: 8

引入可解释物理控制到视频生成，通过物理监督微调与VLM奖励优化，提升物理一致性。

👨‍🔬 Sriram Narayanan, Ziyu Jiang, Srinivasa Narasimhan, Manmohan Chandraker

Crab: A Semantics-Aware Checkpoint/Restore Runtime for Agent Sandboxes

学术论文 ArXiv 重要度: 8

语义感知检查点运行时，通过eBPF分析智能体动作，减少87%检查点流量，保持正确性。

👨‍🔬 Tianyuan Wu, Chaokun Chang, Lunxi Cao, Wei Gao, Wei Wang

Latent Adversarial Detection: Adaptive Probing of LLM Activations for Multi-Turn Attack Detection

学术论文 ArXiv 重要度: 8

利用激活轨迹检测多轮提示注入，将检测率从76.2%提升至93.8%，跨模型可复现。

👨‍🔬 Prashant Kulkarni

PRISM: Pre-alignment via Black-box On-policy Distillation for Multimodal Reinforcement Learning

学术论文 ArXiv 重要度: 8

提出多模态强化学习预对齐流程，通过黑盒蒸馏缓解分布漂移，提升下游RL性能4-6点。

👨‍🔬 Sudong Wang, Weiquan Huang, Xiaomin Yu, Zuhao Yang, Hehai Lin, Keming Wu, Chaojun Xiao, Chen Chen, Wenxuan Wang, Beier Zhu, Yunjian Zhang, Chengwei Qin

LLM as Clinical Graph Structure Refiner: Enhancing Representation Learning in EEG Seizure Diagnosis

学术论文 ArXiv 重要度: 7

利用大语言模型优化脑电图图结构，去除冗余边，显著提升癫痫检测准确率与图可解释性。

👨‍🔬 Lincan Li, Zheng Chen, Yushun Dong

FlexiTac: A Low-Cost, Open-Source, Scalable Tactile Sensing Solution for Robotic Systems

学术论文 ArXiv 重要度: 7

低成本开源触觉传感器，支持密集触觉信号与实时控制，可集成多种机器人平台。

👨‍🔬 Binghao Huang, Yunzhu Li

Do Sparse Autoencoders Capture Concept Manifolds?

学术论文 ArXiv 重要度: 7

理论分析稀疏自编码器捕获概念流形的两种方式，发现实际处于稀释状态，需后处理发现群组。

👨‍🔬 Usha Bhalla, Thomas Fel, Can Rager, Sheridan Feucht, Tal Haklay, Daniel Wurgaft, Siddharth Boppana, Matthew Kowal, Vasudev Shyam, Jack Merullo, Atticus Geiger, Ekdeep Singh Lubana

Normativity and Productivism: Ableist Intelligence? A Degrowth Analysis of AI Sign Language Translation Tools for Deaf People

学术论文 ArXiv 重要度: 6

批判性分析手语翻译AI中的健全主义偏见，指出技术标准化忽视手语文化，造成异化。

👨‍🔬 Nina Seron-Abouelfadil, Poppy Fynes

🤖 AI资讯日报

📊 今日趋势总结

MIT Non-AI License

Why Boring Businesses Outlast AI Hype Cycles

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

Ask HN: Anyone concerned about NYC Local Law 144?

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

The AI Crackpot Index

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

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

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

Show HN: Startup Raising capital through Book Sales

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

Bioinformatician

Synthetic Computers at Scale for Long-Horizon Productivity Simulation

Intern-Atlas: A Methodological Evolution Graph as Research Infrastructure for AI Scientists

Claw-Eval-Live: A Live Agent Benchmark for Evolving Real-World Workflows

Computing Equilibrium beyond Unilateral Deviation

PhyCo: Learning Controllable Physical Priors for Generative Motion

Crab: A Semantics-Aware Checkpoint/Restore Runtime for Agent Sandboxes

Latent Adversarial Detection: Adaptive Probing of LLM Activations for Multi-Turn Attack Detection

PRISM: Pre-alignment via Black-box On-policy Distillation for Multimodal Reinforcement Learning

LLM as Clinical Graph Structure Refiner: Enhancing Representation Learning in EEG Seizure Diagnosis

FlexiTac: A Low-Cost, Open-Source, Scalable Tactile Sensing Solution for Robotic Systems

Do Sparse Autoencoders Capture Concept Manifolds?

Normativity and Productivism: Ableist Intelligence? A Degrowth Analysis of AI Sign Language Translation Tools for Deaf People

📅 历史日报目录