AWhat changed
piKL anchors search to imitation-learned policies through KL-regularized utilities; DiL-piKL distributes regularization strengths and RL-DiL-piKL trains Diplodocus for human-compatible no-press Diplomacy.
Scientific scope
piKL anchors search to imitation-learned policies through KL-regularized utilities; DiL-piKL distributes regularization strengths and RL-DiL-piKL trains Diplodocus for human-compatible no-press Diplomacy. Related planning and learning extensions are one contribution family, with source-specific bylines. Regularized regret bounds are not universal unregularized Nash guarantees: two-player zero-sum equilibrium statements require their stated assumptions, positive-support anchor/log constants and appropriate regularization. Seven-player tournament strength is empirical, and no-press omits natural-language negotiation.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Related planning and learning extensions are one contribution family, with source-specific bylines. Regularized regret bounds are not universal unregularized Nash guarantees: two-player zero-sum equilibrium statements require their stated assumptions, positive-support anchor/log constants and appropriate regularization. Seven-player tournament strength is empirical, and no-press omits natural-language negotiation.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- Modeling Strong and Human-Like Gameplay with KL-Regularized Search
- Mastering the Game of No-Press Diplomacy via Human-Regularized Reinforcement Learning and Planning
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction