Mastering Board Games by External and Internal Planning with Language Models

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Schultz, John, Adamek, Jakub, Jusup, Matej, Lanctot, Marc, Kaisers, Michael, Perrin, Sarah, Hennes, Daniel, Shar, Jeremy, Lewis, Cannada, Ruoss, Anian, Zahavy, Tom, Veličković, Petar, Prince, Laurel, Singh, Satinder, Malmi, Eric, Tomašev, Nenad
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912388669767680
author Schultz, John
Adamek, Jakub
Jusup, Matej
Lanctot, Marc
Kaisers, Michael
Perrin, Sarah
Hennes, Daniel
Shar, Jeremy
Lewis, Cannada
Ruoss, Anian
Zahavy, Tom
Veličković, Petar
Prince, Laurel
Singh, Satinder
Malmi, Eric
Tomašev, Nenad
author_facet Schultz, John
Adamek, Jakub
Jusup, Matej
Lanctot, Marc
Kaisers, Michael
Perrin, Sarah
Hennes, Daniel
Shar, Jeremy
Lewis, Cannada
Ruoss, Anian
Zahavy, Tom
Veličković, Petar
Prince, Laurel
Singh, Satinder
Malmi, Eric
Tomašev, Nenad
contents Advancing planning and reasoning capabilities of Large Language Models (LLMs) is one of the key prerequisites towards unlocking their potential for performing reliably in complex and impactful domains. In this paper, we aim to demonstrate this across board games (Chess, Fischer Random / Chess960, Connect Four, and Hex), and we show that search-based planning can yield significant improvements in LLM game-playing strength. We introduce, compare and contrast two major approaches: In external search, the model guides Monte Carlo Tree Search (MCTS) rollouts and evaluations without calls to an external game engine, and in internal search, the model is trained to generate in-context a linearized tree of search and a resulting final choice. Both build on a language model pre-trained on relevant domain knowledge, reliably capturing the transition and value functions in the respective environments, with minimal hallucinations. We evaluate our LLM search implementations against game-specific state-of-the-art engines, showcasing substantial improvements in strength over the base model, and reaching Grandmaster-level performance in chess while operating closer to the human search budget. Our proposed approach, combining search with domain knowledge, is not specific to board games, hinting at more general future applications.
format Preprint
id arxiv_https___arxiv_org_abs_2412_12119
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Mastering Board Games by External and Internal Planning with Language Models
Schultz, John
Adamek, Jakub
Jusup, Matej
Lanctot, Marc
Kaisers, Michael
Perrin, Sarah
Hennes, Daniel
Shar, Jeremy
Lewis, Cannada
Ruoss, Anian
Zahavy, Tom
Veličković, Petar
Prince, Laurel
Singh, Satinder
Malmi, Eric
Tomašev, Nenad
Artificial Intelligence
Computation and Language
Machine Learning
Advancing planning and reasoning capabilities of Large Language Models (LLMs) is one of the key prerequisites towards unlocking their potential for performing reliably in complex and impactful domains. In this paper, we aim to demonstrate this across board games (Chess, Fischer Random / Chess960, Connect Four, and Hex), and we show that search-based planning can yield significant improvements in LLM game-playing strength. We introduce, compare and contrast two major approaches: In external search, the model guides Monte Carlo Tree Search (MCTS) rollouts and evaluations without calls to an external game engine, and in internal search, the model is trained to generate in-context a linearized tree of search and a resulting final choice. Both build on a language model pre-trained on relevant domain knowledge, reliably capturing the transition and value functions in the respective environments, with minimal hallucinations. We evaluate our LLM search implementations against game-specific state-of-the-art engines, showcasing substantial improvements in strength over the base model, and reaching Grandmaster-level performance in chess while operating closer to the human search budget. Our proposed approach, combining search with domain knowledge, is not specific to board games, hinting at more general future applications.
title Mastering Board Games by External and Internal Planning with Language Models
topic Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2412.12119