Splendor AI
Training competing agents to play a three-player version of Splendor through self-play and tournament evaluation.
DISCIPLINES & TOOLS
The idea
This project trains agents to play a three-player version of Splendor.
Approach
The work includes implementing the game environment, training multiple competing agents with self-play, and evaluating their performance in tournaments. Python and PyTorch form the core of the implementation.
Evaluation
Experimental setup, results, and visualizations will be added when available.
Further materials
The repository, demonstrations, and relevant references can be linked here when they are ready.