Student of Games: A unified learning algorithm for both perfect and imperfect information games

https://doi.org/10.1126/sciadv.adg3256

Guided search, self-play learning, and game-theoretic reasoning in one sound algorithm; strong in chess and Go, beat the strongest open heads-up no-limit agent, and beat the prior best Scotland Yard agent.

Regret minimization and search for imperfect-information games