Metraipolis › archive › dc_a4a5dca7b8ca
Beyond Search-Imitation: Prior-Directed Exploration for Searchless Chess — arXiv:2608.27757
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.27757
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T06:32:32.218Z
- SHA-256 of the archived text
b9e9e9a460bde98b4c3511e102e89f8836e166a1d6b80509f6b5f4f5fa31f421- Stable URL of this page
https://metraipolis.com/documento/dc_a4a5dca7b8ca
How to cite
Beyond Search-Imitation: Prior-Directed Exploration for Searchless Chess — arXiv:2608.27757 — arXiv (en), https://arxiv.org/abs/2608.27757 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a4a5dca7b8ca (Metraipolis archive dc_a4a5dca7b8ca, sha256 b9e9e9a460bde98b4c3511e102e89f8836e166a1d6b80509f6b5f4f5fa31f421)
BibTeX
@misc{dca4a5dca7b8ca,
title = {Beyond Search-Imitation: Prior-Directed Exploration for Searchless Chess — arXiv:2608.27757},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.27757},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a4a5dca7b8ca; sha256 b9e9e9a460bde98b4c3511e102e89f8836e166a1d6b80509f6b5f4f5fa31f421},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.27757. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Searchless chess networks reach human master strength from a single forward pass by imitating a stronger teacher: the strongest, Leela Chess Zero's (Lc0) released Chessformer, distills the visit counts of an AlphaZero-style Monte Carlo Tree Search (MCTS). Imitating a search is a poor proxy for playing without one, so we fine-tune for single-pass strength with self-play reinforcement learning (RL). Its exploration is usually supplied by an entropy bonus, the reverse Kullback-Leibler (KL) divergence to uniform. We replace it with a forward, mass-covering KL toward the network's own MCTS prior (prior-directed exploration), so exploration covers the moves the prior judges promising, and pair it with an entropy-adaptive sampling temperature, set by the value head's outcome uncertainty, that sharpens once a position is decided. In about two thousand steps it raises puzzle accuracy from 93.9% to 94.9% on a 100,000-puzzle suite and mate-in-four accuracy from 77% to 81% while holding searchless strength at or slightly above the base. Measuring tactical accuracy and playing strength together across a matched-compute sweep, we find the two dissociate: accuracy gains fall in a one-point band while ratings straddle the base, and a control fine-tuned on puzzles alone posts the study's largest tactical gains while shedding roughly 260 Elo; a better puzzle-solver is not thereby a stronger player. Distribution-level measurements show what anchoring buys: without a regularizer self-play collapses onto a single line of play, and the puzzles newly solved are the near misses whose winning move the prior kept alive. The forward-KL prior tops the rating ladder, statistically tied with a reverse-KL anchor that concentrates twice as hard and drops the hardest solutions the mass-covering prior keeps in support. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Machine Learning (stat.ML). Pagina del lavoro: https://arxiv.org/abs/2608.27757 — 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 2026-10-17T06:15:12.705Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).