English(EN) Which Successes Make Good References? Action-Sequence Alignment for Multi-Turn On-Policy Distillation

新方法通过动作序列对齐改进AI代理训练

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-11 05:00

研究人员开发了一种新方法，用于改进多轮工具使用代理的在轨蒸馏（OPD）。该技术将成功参考选择视为一个自由目标最优传输问题，旨在将失败的轨迹与最兼容的成功轨迹进行匹配。通过使用动作序列对齐，特别是最长公共子序列（LCS）算法，作为兼容性的代理，该方法识别出共享的动作骨干，同时允许绕行。在ALFWorld和$\tau^3$-bench上的实验表明，LCS比无参考OPD的性能有所提高，并且在新领域中具有可迁移性。 AI

影响这项研究可能导致更有效的AI代理工具使用训练方法，从而提高它们在复杂、多轮任务中的性能。

排序理由该集群包含一篇详细介绍AI代理训练新方法的论文。[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: 34 / 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AI代理训练新方法的论文。[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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Mufei Li, Nikki Lijing Kuang, Zhuoer Wang, Samson Koelle, Jie Yuan, Manqing Mao, Hong Wang, Yanjun Lin, James Feng, Daniel Edmiston, Pan Li, Wei Niu · 2026-11-11 05:00

哪些成功案例可作参考？多轮on-policy蒸馏的动作序列对齐

arXiv:2611.13135v1 Announce Type: cross Abstract: On-policy distillation (OPD) reduces train-inference distribution mismatch by supervising a student on its own trajectories. For multi-turn tool-use agents, conditioning the teacher on a validated same-task successful trajectory a…

本文如何被排名

报道来源 [1]

哪些成功案例可作参考？多轮on-policy蒸馏的动作序列对齐

相关实体

相关话题