Metraipolis › archive › dc_04de3cc3360b
Dependency-Aware Trajectory Refinement for Efficient Multi-Turn Agent Fine-Tuning — arXiv:2609.18417
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.18417
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T13:49:27.100Z
- SHA-256 of the archived text
8a02c922549a274adb3ed825a08d853eda7ef87ad0f1ebc3e3d4f41e399d8d03- Stable URL of this page
https://metraipolis.com/documento/dc_04de3cc3360b
How to cite
Dependency-Aware Trajectory Refinement for Efficient Multi-Turn Agent Fine-Tuning — arXiv:2609.18417 — arXiv (en), https://arxiv.org/abs/2609.18417 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_04de3cc3360b (Metraipolis archive dc_04de3cc3360b, sha256 8a02c922549a274adb3ed825a08d853eda7ef87ad0f1ebc3e3d4f41e399d8d03)
BibTeX
@misc{dc04de3cc3360b,
title = {Dependency-Aware Trajectory Refinement for Efficient Multi-Turn Agent Fine-Tuning — arXiv:2609.18417},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.18417},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_04de3cc3360b; sha256 8a02c922549a274adb3ed825a08d853eda7ef87ad0f1ebc3e3d4f41e399d8d03},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.18417. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Multi-turn agent trajectories often contain redundant rounds (failed tool calls, parallel sub-queries, verification-only steps) that inflate both training and inference cost. We propose viewing each trajectory as a \emph{round-level dependency DAG} that exposes which rounds are globally load-bearing for the final answer, and fine-tune agents on trajectories refined through this DAG. Given an LLM-annotated DAG, these edits are deterministic and interpretable, with optional rephrasing. Models trained on these refined trajectories consistently outperform those trained on the original trajectories at lower inference cost. Specifically, across four multi-modal QA benchmarks, our refinements improve downstream accuracy by up to $1.7$\,pp over vanilla SFT (and $5.7$\,pp over an LLM-deletion baseline) while reducing per-sample inference messages by up to approximately $40\%$ and inference tokens by up to approximately $48\%$, translating to substantial savings in compute and serving cost. Code is available.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-17 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.18417 — 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 2026-10-15T13:48:24.318Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).