Metraipolis › archive › dc_fc80a1096df8
ReSI-Math: Exploring Recoverable Capability Frontiers for Efficient Reasoning Improvement — arXiv:2611.04787
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.04787
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-04T18:37:19.035Z
- SHA-256 of the archived text
fe8d3629e87cdb2ae8bc12740febbd922551b5005eb2313711194ee12d2d2bd3- Stable URL of this page
https://metraipolis.com/documento/dc_fc80a1096df8
How to cite
ReSI-Math: Exploring Recoverable Capability Frontiers for Efficient Reasoning Improvement — arXiv:2611.04787 — arXiv (en), https://arxiv.org/abs/2611.04787 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fc80a1096df8 (Metraipolis archive dc_fc80a1096df8, sha256 fe8d3629e87cdb2ae8bc12740febbd922551b5005eb2313711194ee12d2d2bd3)
BibTeX
@misc{dcfc80a1096df8,
title = {ReSI-Math: Exploring Recoverable Capability Frontiers for Efficient Reasoning Improvement — arXiv:2611.04787},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.04787},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fc80a1096df8; sha256 fe8d3629e87cdb2ae8bc12740febbd922551b5005eb2313711194ee12d2d2bd3},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04787. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reinforcement learning with verifiable rewards (RLVR) has become an effective approach for improving mathematical reasoning in large language models. However, a fundamental question remains: when a model fails to solve a problem, does it truly lack the underlying capability? In this work, we reveal that direct failure does not necessarily imply inability. Some reasoning failures can be recovered by exploring the model's own trajectories, exposing a recoverable capability frontier: a policy-dependent region where solutions are not reliably generated through direct decoding but can be elicited through trajectory-level search. Based on this observation, we introduce ReSI-Math, a framework that explores recoverable capability frontiers and converts recovered capabilities into data-efficient RLVR supervision. ReSI-Math uses Self-Judge-guided evolutionary search to identify recoverable failures under a fixed policy and adaptively selects frontier problems for sequential RLVR as the policy evolves. On Qwen3-VL-8B, correct-incorrect trajectory crossover produces 272 verifier-confirmed correct offspring from 621 attempts (43.8%). After three sequential one-shot RLVR stages, the MATH500 and WE-MATH average improves from 78.71 to 80.16. Under matched single-problem RLVR, random selection reaches only 53.31 versus 54.16 for frontier-guided Q1. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2611.04787 — 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 2026-11-04T18:32:37.464Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).