1
阿里千问发布 QwenGyre，面向超长程智能体任务的在线强化学习框架
IT之家 10 月 12 日消息，阿里千问今日正式发布 QwenGyre，这是一款面向长程智能体任务的在线强化学习框架。 据官方介绍，当智能体承担复杂工作，一次执行往往持续数小时，模型与环境之间有数百轮相互依赖的交互，累计处理百万级 token。QwenGyre 通过弹性调度与轨迹处理， 让这类超长…
2
3
2026 年诺贝尔化学奖公布：二人因解开不对称有机合成难题获奖，其中一位 96 岁高龄
IT之家 10 月 7 日消息，2026 年诺贝尔化学奖今日正式公布，瑞典皇家科学院决定将 2026 年诺贝尔化学奖授予 亨利 ·B· 卡甘（Henri B. Kagan） 和 硖合宪三（Kenso Soai） ，“表彰他们在不对称有机合成中发现非线性效应和自催化现象。” 有些分子，例如氨基酸，以两…
4
罗福莉沉寂半年官宣小米强化学习！直播新模型训练过程，一小时烧3万美元
奖励曲线、显卡故障全公开
5
一张GPU跑10万原子！分子之心用AI把化学反应“拍”成了电影
打破分子模拟“不可能三角”
6
机器人不能停下来等模型：星尘发布 SmoothRL，让在线强化学习跟上大模型的异步推理
星尘智能（Astribot） 基座模型团队发布能异步执行的在线强化学习框架 SmoothRL
7
前字节强化学习专家孙鹏博士加盟星尘智能，完善Physical AI全栈技术布局
9月2日，前字节跳动强化学习专家、前腾讯Robotics X智能体中⼼负责⼈孙鹏博⼠正式加入星尘智能
8
Learning Generalizable Behaviors for Terminal Agents
用强化学习让大模型在终端环境中学会通用行为，打开日常自动化新可能。
arXiv:2608.22631v1 Announce Type: new Abstract: Terminal agents are a compelling application of large language models (LLMs), with the potential to in…
9
RetroDFM-R: Reasoning-Driven Retrosynthesis Prediction with Large Language Models via Reinforcement Learning
用强化学习让大模型学会「讲道理」，逆合成预测从黑箱猜测走向可解释推理——AI驱动分子设计又进一步。
arXiv:2507.17448v2 Announce Type: replace-cross Abstract: Retrosynthetic planning is a cornerstone of organic synthesis and drug discovery. Yet existi…
10
Reinforcing Multi-Turn Reasoning in LLM Agents via Fine-Grained Reward Structure and Credit Assignment
多轮推理强化学习新方案，细粒度奖励结构直击信用分配难题，提升LLM代理决策可靠性。
arXiv:2505.11821v3 Announce Type: replace Abstract: Reinforcement Learning (RL) approaches have been wildly used to enhance the reasoning capabilities…
11
Show HN: Learn Leap, an AI tutor that teaches from your own material
AI tutor 直接从你的资料出题，读书科研的好帮手，告别泛泛而谈的通用问答
I built Learn Leap because I found myself constantly asking ChatGPT questions while reading research papers. I wanted something that already understoo…
12
Credit Without Ground Truth: Auditing Step-Level Credit Assignment in LLM Agents Against Executed Replay
一步错步步错？实测三类信用信号在LLM智能体训练中均无法锁定关键步骤，揭示因果归因盲区。
arXiv:2608.19760v1 Announce Type: new Abstract: Audited against causal ground truth from executed replay in a single-agent tool environment (ALFWorld)…
13
Future Policy Approximation for Offline Reinforcement Learning in LLM Reasoning
离线强化学习遇上LLM推理，未来策略近似方法带来新突破，AI研究者不容错过。
arXiv:2509.19893v3 Announce Type: replace Abstract: Reinforcement learning (RL) has emerged as a key driver of post-training for complex reasoning in …
14
Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Requirements
只用极少GPU就能微调长时程LLM智能体，破解长程推理训练算力困局的新方案。
arXiv:2608.17310v1 Announce Type: new Abstract: Reinforcement Learning (RL) has been promising in single-turn LLM fine-tuning. However, long-horizon a…
15
Effective Personalized AI Tutors via LLM-Guided Reinforcement Learning
大模型结合强化学习，打造精准适配每位学生的AI导师，个性化教育迎来新突破。
arXiv:2608.16907v1 Announce Type: cross Abstract: Generative AI (GenAI) is rapidly reshaping education by unlocking the potential for personalized tut…
16
Agent Lightning v1.0: Towards Harnessed Agentic RL
突破性智能体强化学习框架，解锁自主决策新范式，AI研究者必读前沿成果。
arXiv:2608.17528v1 Announce Type: new Abstract: Modern agents operate inside agent harnesses that manage tools, context, and control flow, making the …
17
Belayer: Efficient Fault Tolerance for LLM Agentic RL Training
应对大模型智能体强化学习训练中的故障难题，为长时训练提供高效容错保障，值得关注。
arXiv:2608.14635v1 Announce Type: cross Abstract: Large language model (LLM) agents are increasingly trained with reinforcement learning in long-horiz…
18
Le Critique: Privileged Value Functions for LLM Reinforcement Learning
给大模型强化学习装上“特权价值函数”，让模型更懂自己表现，训练效率有望大幅提升。
arXiv:2608.16739v1 Announce Type: new Abstract: Reinforcement learning algorithms for Large Language Models (LLMs) are largely distinguished by their …
19
TRCA: Transition-wise Rubric Credit Assignment for Long-horizon LLM Agents
面向长时程智能体，TRCA用过渡级评分信用分配破解稀疏奖励难题，值得关注。
arXiv:2608.16156v1 Announce Type: new Abstract: Long-horizon large language model (LLM) agents are typically optimized with sparse terminal outcomes, …
20
Show HN: Personal AI tutor that builds and probes your understanding
用AI定制从入门到进阶的课程，再通过提问检测你的理解，开源可自部署。
I used to learn from books and courses. But I kept realising I hadn't built a real mental model. I always needed an LLM on the side to fill in what th…