A Doudizhu reinforcement learning AI

This repository is the accompanying code for the paper CFVFP. This paper presents a new algorithm for solving incomplete information games - CFVFP, which has better results than MCCFR in some games. The paper has been accepted by NeurIPS 2024.

Experimenting with intelligent agents utilizing a variety of different algorithmic methods to develop an expert level bot for the card game of Schnapsen.

Awesome Game AI materials of Multi-Agent Reinforcement Learning

[DESAFIO] ¿Cómo hacer que una computadora ( inteligencia artificial ) juegue mejor que un humano al truco? El Truco es un juego de cartas Argentino popularmente jugado en el todo el país y el cono sur. Es un juego de estrategia competitivo basado en turnos, de estados finitos e información incompleta, lo cúal quiere decir que los jugadores basar…

[NeurIPS 2022] PerfectDou: Dominating DouDizhu with Perfect Information Distillation

Counterfactual Regret Minimization Game solution methods for Julia

Deriving an optimal Blackjack strategy using Tammelin's counterfactual regret minimization algorithm (CFR+) and modern C++.

Series of tools to evaluate and examine the game Dark Hex

C++ implementations of Counterfactual Regret Minimization and Monte Carlo CFR

Python library of imperfect information games, and self-play and exploitative algorithms

Analyzing Betting Game and its applications specially in poker

Opponent exploitation in imperfect game using Neural Networks.

I made 2 different AIs (a neural network and a Q-learning algithm) play an imperfect information game and observe the strategys that both players used.

Scalable Implementation of Deep CFR and Single Deep CFR