English(EN) SCOPE: Selective Critical-Step On-Policy Self-Distillation for Agentic Reinforcement Learning

新的SCOPE框架增强了LLM代理在强化学习中的性能

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-31 04:00

研究人员开发了SCOPE（Selective Critical-Step On-Policy Self-Distillation），一个新颖的框架，旨在提高LLM代理在强化学习任务中的性能。该方法通过分析代理轨迹、生成特定步骤的技能，并仅在选定的错误步骤应用蒸馏来优化学习过程。在ALFWorld和WebShop基准上的实验显示出显著的改进，Qwen2.5-3B等模型取得了高成功率，并在未见过任务上超越了SEED和GRPO等现有方法。 AI

影响这项研究引入了一种新颖的蒸馏技术，有望为复杂、长周期的任务带来更高效、更强大的LLM代理。

排序理由该集群包含一篇详细介绍代理强化学习新框架的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍代理强化学习新框架的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: 3 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Qirui Liu, Wei Cao, Yu Mi, Yichen Sun, Xuekun Zhang, Lei Liu, Yue Shen, Zhixuan Chu · 2026-10-31 04:00

SCOPE：用于智能体强化学习的选择性关键步骤策略内自蒸馏

arXiv:2610.38391v1 Announce Type: new Abstract: Outcome-based reinforcement learning provides coarse supervision for long-horizon LLM agents: failed trajectories may contain correct steps, while successful trajectories can contain errors. Trajectory-level skills can also shift te…

本文如何被排名

报道来源 [1]

SCOPE：用于智能体强化学习的选择性关键步骤策略内自蒸馏

相关实体

相关话题