Metraipolis › archive › dc_e71cfd1cbea8
Adaptive Mixing of Policies from Searching and Policies from Learning — arXiv:2608.15700
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.15700
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T17:26:10.404Z
- SHA-256 of the archived text
64a09320f51819b3ca61018c016ab968971716febb2b3e020dbd30b78f8f6e3e- Stable URL of this page
https://metraipolis.com/documento/dc_e71cfd1cbea8
How to cite
Adaptive Mixing of Policies from Searching and Policies from Learning — arXiv:2608.15700 — arXiv (en), https://arxiv.org/abs/2608.15700 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e71cfd1cbea8 (Metraipolis archive dc_e71cfd1cbea8, sha256 64a09320f51819b3ca61018c016ab968971716febb2b3e020dbd30b78f8f6e3e)
BibTeX
@misc{dce71cfd1cbea8,
title = {Adaptive Mixing of Policies from Searching and Policies from Learning — arXiv:2608.15700},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.15700},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e71cfd1cbea8; sha256 64a09320f51819b3ca61018c016ab968971716febb2b3e020dbd30b78f8f6e3e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.15700. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Background: Distillation of training targets generated thru search/planning has proven useful in reinforcement learning, but search can take exceedingly long. Objectives: Rather than perform search to the same depth every time (typically at a fixed period of steps), reduce the search depth proportionally to the quality of the policy network priors. Methods: We describe Flexer, an architecture that, for each step, mixes the policy from a neural network and the policy from Monte Carlo tree search. The mixing factor favors the MCTS policy as the policy imitation error of the network and the environment models' variance increases. Results: Flexer outperforms a version of AlphaZero (and DQN and ADP) for some experiments on three toy symbolic problems. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.15700 — 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 2026-10-18T16:29:13.823Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).