Verified65% confidenceEventExact time
Libratus, developed by Carnegie Mellon University in 2017, defeated top professional players in no-limit Texas Hold'em poker, with its core algorithm being Counterfactual Regret Minimization (CFR).
3
Sources
65%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
How Does AI Conquer Imperfect Information Games? A Deep Dive into Tactico
redditr/reinforcementlearning7/11/2026
Related Claims
UnverifiedCounterfactual regret minimization (CFR) is an iterative algorithm that approximates a Nash equilibrium by minimizing counterfactual regret and has become the mainstream solution method for games with incomplete information such as Texas Hold'em.68% similarUnverifiedMCTS does not rely on domain-specific evaluation functions but instead evaluates position value through large numbers of simulations, so in games with incomplete information such as poker it is often used in combination with techniques such as opponent modeling and counterfactual regret minimization (CFR).55% similarUnverifiedThe tournament sampling design of SynthID-Text has a deep connection with the Gumbel-max trick in probability theory.54% similarUnverifiedCounterfactual Regret Minimization (CFR) was proposed by Zinkevich et al. in 2007 and was originally used to solve Nash equilibria in large-scale imperfect information games.54% similarUnverifiedTypical scenarios of abstraction leakage include distributed training getting stuck at all-reduce, mixed precision causing gradient overflow, and custom operators not being supported.52% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/491879API
curl https://kongchang.com/api/v1/knowledge/claims/491879MCP
get_claim(id=491879)