Metraipolis › archive › dc_e9ef800b93cd

Trajectory2Gym: Decision-Grounded Compositional Interaction Synthesis for Tool-Use Agents — arXiv:2610.33568

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.33568
License (as archived)
CC0-1.0
Archived
2026-10-27T08:11:22.555Z
SHA-256 of the archived text
065795b60b7feda7cd0b2539dd57ac7cd385301db4de1fc1f739443a97bf504c
Stable URL of this page
https://metraipolis.com/documento/dc_e9ef800b93cd

How to cite

Trajectory2Gym: Decision-Grounded Compositional Interaction Synthesis for Tool-Use Agents — arXiv:2610.33568 — arXiv (en), https://arxiv.org/abs/2610.33568 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e9ef800b93cd (Metraipolis archive dc_e9ef800b93cd, sha256 065795b60b7feda7cd0b2539dd57ac7cd385301db4de1fc1f739443a97bf504c)
BibTeX
@misc{dce9ef800b93cd,
  title = {Trajectory2Gym: Decision-Grounded Compositional Interaction Synthesis for Tool-Use Agents — arXiv:2610.33568},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.33568},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e9ef800b93cd; sha256 065795b60b7feda7cd0b2539dd57ac7cd385301db4de1fc1f739443a97bf504c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.33568. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

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.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.33568 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-27T07:56:00.202Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).