Meet New AI System That Can Play Poker and Chess
DeepMind, an AI lab powered by Google parent company Alphabet (NASDAQ: GOOGL), has created an AI system that can play games.
Named the Player of Games, the system can perform excellently at perfect information games like chess and the Chinese board game, Go, and imperfect information games like poker.
According to the lab, the system can reason about others’ goals and motivations through the task-solving skills it gets from games, including contract negotiations, route planning around congestion, and interacting with customers.
With these skills, the Player of Games could lay the framework for AI that can be well trusted and successfully work with others in the future.
DeepMind is reputable for creating game-playing AI systems. The lab believes that although games lack a commercial application, they are great for learning reasoning and cognitive abilities.
Therefore, they are vital in AI building. The knowledge of games for AI has also led to the production of self-learning AI that can power computer vision, self-driving cars, and natural language processing.
Other game-playing systems in the lab’s portfolio include the chess-winning AlphaZero and StarCraft II-besting AlphaStar.
[Player of Games] learns to play [games] from scratch, simply by repeatedly playing the game in self-play.