Metraipolis › archive › dc_ef98a371be01

Beyond Verified Answers: Solver-Informed Self-Distillation for Bootstrapping Operations Research Language Models — arXiv:2609.09957

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.09957
License (as archived)
CC0-1.0
Archived
2026-10-16T05:24:34.547Z
SHA-256 of the archived text
2be025db71352adf80baaf86cf0a852404871795efb3834bb1f92aecd531b119
Stable URL of this page
https://metraipolis.com/documento/dc_ef98a371be01

How to cite

Beyond Verified Answers: Solver-Informed Self-Distillation for Bootstrapping Operations Research Language Models — arXiv:2609.09957 — arXiv (en), https://arxiv.org/abs/2609.09957 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ef98a371be01 (Metraipolis archive dc_ef98a371be01, sha256 2be025db71352adf80baaf86cf0a852404871795efb3834bb1f92aecd531b119)
BibTeX
@misc{dcef98a371be01,
  title = {Beyond Verified Answers: Solver-Informed Self-Distillation for Bootstrapping Operations Research Language Models — arXiv:2609.09957},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.09957},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ef98a371be01; sha256 2be025db71352adf80baaf86cf0a852404871795efb3834bb1f92aecd531b119},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.09957. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Modern large language models (LLMs) can translate natural-language descriptions into operations research (OR) formulations. Post-training techniques including reinforcement learning and on-policy self-distillation have further improved this capability. However, three limitations remain in training LLMs for OR formulations. First, training commonly relies on synthetic formulations validated by human experts or stronger models, constraining scalable supervision. Second, credit assignment is either coarse or costly: outcome rewards score an entire trajectory without locating the responsible modeling decision, whereas process-level supervision requires an additional evaluator. Third, privileged self-distillation can induce style mismatch by using solver context unavailable at deployment. We find that a model can improve from solver-artifact feedback generated by its own rollouts, making self-distillation a practical, evaluator-free source of dense supervision. Therefore, we propose SOLID: Solver-Informed On-Policy LearnIng through Self-Distillation, a novel framework for self-improving OR language models without verified answers or external evaluators. SOLID executes candidate programs from multiple rollouts, clusters their objectives, and selects a majority-group artifact as a pseudo-reference. The model then performs updates using group-relative advantages and dense self-supervision signals. Across multiple OR benchmarks, SOLID improves solution accuracy for both general-purpose and OR-tuned models over outcome-only group-relative training. These results show that solver artifacts can support scalable self-improvement without trusted answers.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-11 in cs.AI).
Categorie arXiv: Optimization and Control (math.OC) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.09957 — 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 2026-10-16T05:17:52.064Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).