ANR Noggins (IRIT)

Entrainer des modèles de joueurs pour extraire des stratégies de jeu et faire de la vérification automatique.

This project aims at building intelligent players that can play according to different objectives, from “winning the game” to “being ethical”, by learning and representing in a symbolic way the strategies for playing games. To do so, agents will first reason about the rules governing the games, and second about prior knowledge about strategies. The agents will then be able to either construct strategies by guiding the exploration of strategic moves, or to adapt and refine existing strategies. The key impact is to build trustworthy interpretable AI players which consider strategies as plain meaningful objects. As game-playing recently became a test-bed for AI innovations, this project aims to inspire future advancements in the field beyond the realm of games.