Senet AI trains a reinforcement learning agent to play Senet , one of the oldest known board games using tabular Q-learning and self-play, with no human demonstrations or hand-coded strategy. Two agents compete over thousands of episodes, learning move selection purely from experience. Includes three trained difficulty levels and a resizable Pygame GUI for playing against the AI or watching AI vs AI matches.
- (credits)
- Anrich Kruger
- (Date)
- 2026
- (clients)
- Anrich Kruger | Artificial intelligence project
