AI资讯日报 - 2025/11/4

👨‍🔬 Danyal Maqbool, Changhee Lee, Zachary Huemann, Samuel D. Church, Matthew E. Larson, Scott B. Perlman, Tomas A. Romero, Joshua D. Warner, Meghan Lubner, Xin Tie, Jameson Merkow, Junjie Hu, Steve Y. Cho, Tyler J. Bradshaw

Best Practices for Biorisk Evaluations on Open-Weight Bio-Foundation Models

学术论文 ArXiv 重要度: 8

评估生物基础模型的双重使用风险，发现当前数据过滤方法效果有限，需更鲁棒的安全策略。

👨‍🔬 Boyi Wei, Zora Che, Nathaniel Li, Udari Madhushani Sehwag, Jasper Götting, Samira Nedungadi, Julian Michael, Summer Yue, Dan Hendrycks, Peter Henderson, Zifan Wang, Seth Donoughe, Mantas Mazeika

Visual Backdoor Attacks on MLLM Embodied Decision Making via Contrastive Trigger Learning

学术论文 ArXiv 重要度: 8

提出BEAT框架，通过对比触发学习在多模态大语言模型 embodied agent中植入视觉后门攻击。

👨‍🔬 Qiusi Zhan, Hyeonjeong Ha, Rui Yang, Sirui Xu, Hanyang Chen, Liang-Yan Gui, Yu-Xiong Wang, Huan Zhang, Heng Ji, Daniel Kang

Interaction as Intelligence Part II: Asynchronous Human-Agent Rollout for Long-Horizon Task Training

学术论文 ArXiv 重要度: 7

提出Apollo框架，通过异步人机交互和动作级过滤，有效训练LLM agent完成长周期专业任务。

👨‍🔬 Dayuan Fu, Yunze Wu, Xiaojie Cai, Lyumanshan Ye, Shijie Xia, Zhen Huang, Weiye Si, Tianze Xu, Jie Sun, Keyu Li, Mohan Jiang, Junfei Wang, Qishuo Hua, Pengrui Lu, Yang Xiao, Pengfei Liu

MolChord: Structure-Sequence Alignment for Protein-Guided Drug Design

学术论文 ArXiv 重要度: 7

提出MolChord方法，整合结构序列对齐和偏好优化，实现蛋白质引导的药物分子设计。

👨‍🔬 Wei Zhang, Zekun Guo, Yingce Xia, Peiran Jin, Shufang Xie, Tao Qin, Xiang-Yang Li

Challenges in Credit Assignment for Multi-Agent Reinforcement Learning in Open Agent Systems

学术论文 ArXiv 重要度: 7

分析开放多智能体系统中信用分配挑战，发现开放性导致信用错误归因和性能下降。

👨‍🔬 Alireza Saleh Abadi, Leen-Kiat Soh

Sketch-to-Layout: Sketch-Guided Multimodal Layout Generation

学术论文 ArXiv 重要度: 6

提出草图引导的多模态布局生成方法，利用合成草图数据训练，提供直观设计体验。

👨‍🔬 Riccardo Brioschi, Aleksandr Alekseev, Emanuele Nevali, Berkay Döner, Omar El Malki, Blagoj Mitrevski, Leandro Kieliger, Mark Collier, Andrii Maksai, Jesse Berent, Claudiu Musat, Efi Kokiopoulou

Community Detection on Model Explanation Graphs for Explainable AI

学术论文 ArXiv 重要度: 6

提出影响力模块框架，通过社区检测发现特征组合模式，提升可解释AI的模型调试能力。

👨‍🔬 Ehsan Moradi

VessShape: Few-shot 2D blood vessel segmentation by leveraging shape priors from synthetic images

学术论文 ArXiv 重要度: 6

提出VessShape方法，利用合成图像形状先验实现血管分割的少样本学习，提升泛化能力。

👨‍🔬 Cesar H. Comin, Wesley N. Galvão

Information-Theoretic Greedy Layer-wise Training for Traffic Sign Recognition

学术论文 ArXiv 重要度: 5

提出信息论逐层贪婪训练方法，避免反向传播，在交通标志识别中达到SGD相当性能。

👨‍🔬 Shuyan Lyu, Zhanzimo Wu, Junliang Du

Validity Is What You Need

学术论文 ArXiv 重要度: 5

提出Agentic AI的现实主义定义，强调应用验证的重要性，认为LLM可被更简单模型替代。

👨‍🔬 Sebastian Benthall, Andrew Clark

🤖 AI资讯日报

📊 今日趋势总结

Why Boring Businesses Outlast AI Hype Cycles

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

The AI Crackpot Index

Ask HN: Anyone concerned about NYC Local Law 144?

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

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

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

Ask HN: Dipping my toes with artificial intelligence and what to expect? (CS)

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

Bioinformatician

Show HN: Startup Raising capital through Book Sales

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

Continuous Autoregressive Language Models

PETAR: Localized Findings Generation with Mask-Aware Vision-Language Modeling for PET Automated Reporting

Best Practices for Biorisk Evaluations on Open-Weight Bio-Foundation Models

Visual Backdoor Attacks on MLLM Embodied Decision Making via Contrastive Trigger Learning

Interaction as Intelligence Part II: Asynchronous Human-Agent Rollout for Long-Horizon Task Training

MolChord: Structure-Sequence Alignment for Protein-Guided Drug Design

Challenges in Credit Assignment for Multi-Agent Reinforcement Learning in Open Agent Systems

Sketch-to-Layout: Sketch-Guided Multimodal Layout Generation

Community Detection on Model Explanation Graphs for Explainable AI

VessShape: Few-shot 2D blood vessel segmentation by leveraging shape priors from synthetic images

Information-Theoretic Greedy Layer-wise Training for Traffic Sign Recognition

Validity Is What You Need

📅 历史日报目录