arXiv cs.CL· Yihao Hu, Zhihao Wen, Yitong Lin, Siyuan Lu, Pan Wang, Wei Wu, Jiang-Ming Yang, Tao Lin·· 17 小时前AI 评分32
Trajectory2Gym：面向工具使用智能体的决策锚定组合交互合成
Trajectory2Gym: Decision-Grounded Compositional Interaction Synthesis for Tool-Use Agents
AI 导读
Trajectory2Gym 通过构建可执行交互并仅从有依据的决策中提取监督信号，在五个 Qwen3 规模上均提升 Tau1、Tau2 和 BFCL，三项基准等权平均提升 16.9-29.8 个百分点。该方法强调组合覆盖与决策锚定两项要求，增益在有状态和多轮交互中最为显著。
正文 · 原文
Abstract:Successful tool-use trajectories are not necessarily useful supervision. Even successful interactions may underrepresent multi-step dependencies or contain actions that are unjustified by the information visible to the agent at decision time. We identify two coupled requirements for effective trajectory supervision: compositional coverage, so that important observation-action dependencies are represented, and decision grounding, so that each supervised action is supported by the learner-visible interaction history. We introduce Trajectory2Gym, which constructs executable interactions around under-supported compositional structures and derives supervision only from grounded decisions in their realized executions. Structural-support selection allocates the resulting experience, while parent-normalized supervised fine-tuning preserves the semantic interaction as the unit of training weight. This reframes trajectory generation as the deliberate construction of learnable interaction experience, rather than the collection of successful trajectories. Across five Qwen3 scales, Trajectory2Gym improves Tau1, Tau2, and BFCL at every tested scale, increasing the equal-weight three-benchmark mean by 16.9-29.8 percentage points. The gains are strongest in stateful and multi-turn interaction, where later decisions depend on earlier observations and feedback. Our results suggest that scaling tool-use agents depends not only on collecting more successful trajectories, but on constructing interactions whose dependencies are represented and whose decisions are learnable from the agent's visible history.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.33568 [cs.CL] |
| (or arXiv:2610.33568v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.33568 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yihao Hu [view email]
[v1]
Sat, 24 Oct 2026 09:38:02 UTC (839 KB)
来源：arXiv cs.CL · arxiv.org