Top 4 Rust reinforcement-learning Projects
-
-
InfluxDB
InfluxDB – Built for High-Performance Time Series Workloads. InfluxDB 3 OSS is now GA. Transform, enrich, and act on time series data directly in the database. Automate critical tasks and eliminate the need to move data externally. Download now.
-
robopoker
Play, learn, solve, and analyze No-Limit Texas Hold Em. Implementation follows from Monte Carlo counter-factual regret minimization over with hierarchical K-means imperfect recall abstractions.
Project mention: Professional Poker Players Know the Optimal Strategy but Don't Always Use It | news.ycombinator.com | 2024-07-19One thing I haven't seen anyone mention yet is that Nash equilibria do not actually exist when you move beyond heads-up into multi-way play. There's strong empirical evidence that solving abstracted games using MCCFR and using real-time depth-limited tree search dominates humans and outperforms all other AI strategies, but these results aren't theoretically sound, unlike in heads-up play.
I've actually been working slowly on https://github.com/krukah/robopoker, an open-source Rust implementation of Pluribus, the SOTA poker AI. What I've found interesting is the difference in how I approach actually playing poker versus how I approach building a solver. Playing the game naturally consists of reasoning about narratives and incorporating information like hand history, play style, live tells. Whereas solving the game is about evaluating tradeoffs between the guarantees of imperfect-information game theory and the constraints of Texas Hold'em, finding a balance between abstract and concrete reasoning.
-
-