Metraipolis › archive › dc_96fcb0715eb4
JoMoHa: Joint Model--Harness Optimization with Reinforcement Learning — arXiv:2611.11499
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.11499
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-10T18:43:25.978Z
- SHA-256 of the archived text
c2181138dafa7d8708a1622ab56cf090d0f9af07e2ae6d1e92a1bb8a99074d5e- Stable URL of this page
https://metraipolis.com/documento/dc_96fcb0715eb4
How to cite
JoMoHa: Joint Model--Harness Optimization with Reinforcement Learning — arXiv:2611.11499 — arXiv (en), https://arxiv.org/abs/2611.11499 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_96fcb0715eb4 (Metraipolis archive dc_96fcb0715eb4, sha256 c2181138dafa7d8708a1622ab56cf090d0f9af07e2ae6d1e92a1bb8a99074d5e)
BibTeX
@misc{dc96fcb0715eb4,
title = {JoMoHa: Joint Model--Harness Optimization with Reinforcement Learning — arXiv:2611.11499},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.11499},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_96fcb0715eb4; sha256 c2181138dafa7d8708a1622ab56cf090d0f9af07e2ae6d1e92a1bb8a99074d5e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.11499. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
The performance of language agents depends jointly on model weights and the harness. Optimizing either alone can be constrained by the other component, limiting overall performance gains. Our analysis of naive alternating training reveals progressive incompatibility between the model and harness, which increasingly interferes with task execution and may also undermine model training, leading to performance degradation. To coordinate the evolution of models and harnesses, we introduce JoMoHa, to our knowledge the first framework to jointly train an executable harness generation policy and a task execution policy through reinforcement learning. JoMoHa outperforms the evaluated weight-only and harness-only baselines and achieves the highest success rates among the evaluated methods on ALFWorld, WebShop, and DBBench, averaging 70.3% and reaching 97.0% on ALFWorld. By learning both task execution and runtime improvement from its own generated trajectories, JoMoHa performs recursive self-improvement (RSI). Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2611.11499 — 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 2026-11-10T18:37:37.644Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).