From Imitation to Interaction: Mastering Game of Schnapsen with Shallow Reinforcement Learning
The paper explores the effectiveness of shallow neural network agents in mastering the card game Schnapsen. It compares a supervised learning agent with a reinforcement learning agent, finding that the latter significantly outperforms the former. The study concludes that reinforcement learning, particularly when combined with deeper lookahead strategies, leads to higher winning rates against strong opponents.
- ▪Shallow neural network agents were tested for their ability to master the game Schnapsen.
- ▪Supervised imitation learning did not generalize well against strong opponents, while reinforcement learning showed better results.
- ▪The best performance was achieved by combining learned value functions with deeper lookahead during gameplay.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.17162 (cs) [Submitted on 16 May 2026] Title:From Imitation to Interaction: Mastering Game of Schnapsen with Shallow Reinforcement Learning Authors:Ján Klačan, Sizhong Zhang View a PDF of the paper titled From Imitation to Interaction: Mastering Game of Schnapsen with Shallow Reinforcement Learning, by J\'an Kla\v{c}an and Sizhong Zhang View PDF HTML (experimental) Abstract:This paper investigates whether shallow neural network agents can master the card game Schnapsen and challenge a strong search-based baseline, RdeepBot, which uses Monte Carlo sampling and lookahead search.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
