Metraipolis › archive › dc_f98a211ca698

CIPO: Counterfactual Imagination Policy Optimization for Adaptive Tool Granularity Selection — arXiv:2610.04991

Source
arXiv
Original
https://arxiv.org/abs/2610.04991
License (as archived)
CC0-1.0
Archived
2026-10-06T05:24:55.896Z
SHA-256 of the archived text
81a9951d61e1688149da640bc4f632ed75d2fdd3f2614e74a874e273d1d67755
Stable URL of this page
https://metraipolis.com/documento/dc_f98a211ca698

How to cite

CIPO: Counterfactual Imagination Policy Optimization for Adaptive Tool Granularity Selection — arXiv:2610.04991 — arXiv, https://arxiv.org/abs/2610.04991 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f98a211ca698 (Metraipolis archive dc_f98a211ca698, sha256 81a9951d61e1688149da640bc4f632ed75d2fdd3f2614e74a874e273d1d67755)
BibTeX
@misc{dcf98a211ca698,
  title = {CIPO: Counterfactual Imagination Policy Optimization for Adaptive Tool Granularity Selection — arXiv:2610.04991},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.04991},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f98a211ca698; sha256 81a9951d61e1688149da640bc4f632ed75d2fdd3f2614e74a874e273d1d67755},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.04991. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language model (LLM) agents solve complex tasks through multi-step interactions with external tools. These interactions often contain recurring local tool sequences. Treating such sequences as composite "Skills" can shorten tool-use trajectories and reduce repeated low-level decisions. However, when atomic tools and composite skills coexist, skill use becomes a policy problem: the agent must decide whether the current state requires atomic fine control or skill-level abstraction. In this paper, we argue that effective skill use should be studied as adaptive tool granularity selection. The most direct training signal for this problem is to compare the consequences of atomic and skill choices available from the same state. Based on this view, we propose CIPO, a Counterfactual Imagination Policy Optimization framework for adaptive tool granularity. CIPO constructs executable skills through budget-constrained mining of successful tool-use trajectories and trains granularity decisions with counterfactual branch rollouts. For each base rollout, CIPO branches at the first eligible granularity decision and replaces the chosen action with a feasible atomic or skill alternative. The paired outcome difference serves as a supplementary reward for policy optimization. Experiments across multiple benchmarks and model backbones show that CIPO improves task success and decision efficiency over baselines. Further analyses show that CIPO learns effective skill use by improving the choice between atomic tools and composite skills based on the current state, without simply increasing skill frequency.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.04991 — 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 2026-10-06T05:24:05.540Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).