AI资讯日报 - 2025/10/23

👨‍🔬 Archana Warrier, Dat Nyugen, Michelangelo Naim, Moksh Jain, Yichao Liang, Karen Schroeder, Cambridge Yang, Joshua B. Tenenbaum, Sebastian Vollmer, Kevin Ellis, Zenna Tavares

AdaSPEC: Selective Knowledge Distillation for Efficient Speculative Decoders

学术论文 ArXiv 重要度: 8

提出AdaSPEC方法，通过选择性知识蒸馏提升推测解码的令牌接受率，显著加速LLM推理。

👨‍🔬 Yuezhou Hu, Jiaxin Guo, Xinyu Feng, Tuo Zhao

SmartSwitch: Advancing LLM Reasoning by Overcoming Underthinking via Promoting Deeper Thought Exploration

学术论文 ArXiv 重要度: 7

提出SmartSwitch推理框架，通过检测和干预浅层思考，引导LLM进行深度思考探索，提升推理质量。

👨‍🔬 Xichen Zhang, Sitong Wu, Haoru Tan, Shaozuo Yu, Yinghao Zhu, Ziyi He, Jiaya Jia

Beyond Reactivity: Measuring Proactive Problem Solving in LLM Agents

学术论文 ArXiv 重要度: 7

提出PROBE评估框架，系统测量LLM代理的主动问题解决能力，发现现有模型在此方面存在局限。

👨‍🔬 Gil Pasternak, Dheeraj Rajagopal, Julia White, Dhruv Atreja, Matthew Thomas, George Hurn-Maloney, Ash Lewis

Learning Affordances at Inference-Time for Vision-Language-Action Models

学术论文 ArXiv 重要度: 7

提出LITEN方法，使VLA模型能在推理时从执行经验中学习，动态调整行为以完成复杂控制任务。

👨‍🔬 Ameesh Shah, William Chen, Adwait Godbole, Federico Mora, Sanjit A. Seshia, Sergey Levine

A Survey on Cache Methods in Diffusion Models: Toward Efficient Multi-Modal Generation

学术论文 ArXiv 重要度: 6

系统综述扩散模型缓存方法，提出统一分类框架，为高效多模态生成提供理论基础。

👨‍🔬 Jiacheng Liu, Xinyu Wang, Yuqi Lin, Zhikai Wang, Peiru Wang, Peiliang Cai, Qinming Zhou, Zhengan Yan, Zexuan Yan, Zhengyi Shi, Chang Zou, Yue Ma, Linfeng Zhang

Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning

学术论文 ArXiv 重要度: 6

提出Memo架构，通过周期性摘要令牌实现记忆的高效创建和检索，提升具身代理在长时任务中的表现。

👨‍🔬 Gunshi Gupta, Karmesh Yadav, Zsolt Kira, Yarin Gal, Rahaf Aljundi

Integrating Transparent Models, LLMs, and Practitioner-in-the-Loop: A Case of Nonprofit Program Evaluation

学术论文 ArXiv 重要度: 5

结合透明模型、LLM和从业者参与，为非营利组织提供准确、可解释且可操作的AI评估方案。

👨‍🔬 Ji Ma, Albert Casella

On Controlled Change: Generative AI's Impact on Professional Authority in Journalism

学术论文 ArXiv 重要度: 4

研究新闻业如何通过受控变革管理AI整合，包括制定指南、实验工具和评估能力限制。

👨‍🔬 Tomás Dodds, Wang Ngai Yeung, Claudia Mellado, Mathias-Felipe de Lima-Santos

Misalignment Bounty: Crowdsourcing AI Agent Misbehavior

学术论文 ArXiv 重要度: 3

通过众包项目收集AI代理不当行为案例，为理解和对齐AI系统提供实证基础。

👨‍🔬 Rustem Turtayev, Natalia Fedorova, Oleg Serikov, Sergey Koldyba, Lev Avagyan, Dmitrii Volkov

🤖 AI资讯日报

📊 今日趋势总结

Why Boring Businesses Outlast AI Hype Cycles

The AI Crackpot Index

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

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

Ask HN: Anyone concerned about NYC Local Law 144?

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"?

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

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

Bioinformatician

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

Show HN: Startup Raising capital through Book Sales

OpenHands/OpenHands

Semantic World Models

Scaf-GRPO: Scaffolded Group Relative Policy Optimization for Enhancing LLM Reasoning

Benchmarking World-Model Learning

AdaSPEC: Selective Knowledge Distillation for Efficient Speculative Decoders

SmartSwitch: Advancing LLM Reasoning by Overcoming Underthinking via Promoting Deeper Thought Exploration

Beyond Reactivity: Measuring Proactive Problem Solving in LLM Agents

Learning Affordances at Inference-Time for Vision-Language-Action Models

A Survey on Cache Methods in Diffusion Models: Toward Efficient Multi-Modal Generation

Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning

Integrating Transparent Models, LLMs, and Practitioner-in-the-Loop: A Case of Nonprofit Program Evaluation

On Controlled Change: Generative AI's Impact on Professional Authority in Journalism

Misalignment Bounty: Crowdsourcing AI Agent Misbehavior

📅 历史日报目录