Metraipolis › archive › dc_6da86eec303c
Alliance of Stochastic Routing and Reinforcement Fusion for Multi-Model Reasoning — arXiv:2610.24054
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.24054
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-20T11:41:02.075Z
- SHA-256 of the archived text
f911ea543fbdd4dd02f66e84e683c0dd3447b313f7e280b9d54043e2b0421896- Stable URL of this page
https://metraipolis.com/documento/dc_6da86eec303c
How to cite
Alliance of Stochastic Routing and Reinforcement Fusion for Multi-Model Reasoning — arXiv:2610.24054 — arXiv (en), https://arxiv.org/abs/2610.24054 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6da86eec303c (Metraipolis archive dc_6da86eec303c, sha256 f911ea543fbdd4dd02f66e84e683c0dd3447b313f7e280b9d54043e2b0421896)
BibTeX
@misc{dc6da86eec303c,
title = {Alliance of Stochastic Routing and Reinforcement Fusion for Multi-Model Reasoning — arXiv:2610.24054},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.24054},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6da86eec303c; sha256 f911ea543fbdd4dd02f66e84e683c0dd3447b313f7e280b9d54043e2b0421896},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.24054. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
This paper studies dynamic model routing under uncertainty in multimodal settings, where a system must select and potentially compose multiple models to answer a query. Standard reinforcement learning approaches for routing often suffer from premature convergence to suboptimal models due to noisy rewards and insufficient exploration. We propose a unified vision-language framework (VL-Scout) for efficient multimodal and multi-step reasoning with turn-based reinforcement fusion. First, VL-Scout combines Bayesian priors, policy gradient optimization, and sequential decision-making to enable structured exploration and adaptive inference. A context-dependent Dirichlet belief model allows for epistemic uncertainty over model quality and is updated online based on observed rewards and context-specific evidence, enabling Thompson-style exploration while serving as a prior for regularization. Second, a routing policy is trained via a KL-regularized REINFORCE objective by maximizing a reward-weighted evidence lower bound, where the policy acts as an amortized posterior, and the Dirichlet belief serves as a prior. Third, we introduce a fusion policy trained with Group Relative Policy Optimization (GRPO) or Reinforcement Learning with Verifiable Rewards (RLVR), which operates over intermediate model rationales and selects between two actions: CONTINUE or ANSWER, terminating with a generated output. We evaluate VL-Scout on four representative multi-modal benchmarks (MMMU, MMMU-Pro, OKVQA, and RealWorldQA), compared against 11 representative routing and policy optimization baselines. Our method provides consistent performance improvements while reducing premature policy collapse. These results highlight the importance of integrating uncertainty-aware priors with both amortized policy learning and sequential fusion for robust and efficient model selection in discrete action spaces Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Multiagent Systems (cs.MA). Pagina del lavoro: https://arxiv.org/abs/2610.24054 — 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 2026-10-20T11:41:00.013Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).