Chapter
Action Abstraction and Information Abstraction in Poker
06:58 - 13:12 (06:13)
The action abstraction technology in the beginning finds the optimal combination of bet sizes which is provably convergent, but not very scalable. Meanwhile, information abstraction concerns the abstraction of what chance does in terms of abstraction in general games, since the top players are so good at hiding tells that it is not worth trying to find them in each other.
Clips
The interview discusses how an AI system could win against human players at the game of poker.
06:58 - 07:37 (00:38)
Summary
The interview discusses how an AI system could win against human players at the game of poker. The guest describes the various stages of the game and how the AI program learnt from its mistakes.
ChapterAction Abstraction and Information Abstraction in Poker
EpisodeTuomas Sandholm: Poker and Game Theory
PodcastLex Fridman Podcast
The top players in poker are so good at hiding tells that it's not worth investing a lot of effort into finding them.
07:37 - 10:32 (02:55)
Summary
The top players in poker are so good at hiding tells that it's not worth investing a lot of effort into finding them. However, tells are still considered critical in human versus human poker games.
ChapterAction Abstraction and Information Abstraction in Poker
EpisodeTuomas Sandholm: Poker and Game Theory
PodcastLex Fridman Podcast
This podcast discusses the issue with finer grained abstractions in poker and the use of an automated action abstraction technology for scalable optimal bet sizes.
10:32 - 13:12 (02:39)
Summary
This podcast discusses the issue with finer grained abstractions in poker and the use of an automated action abstraction technology for scalable optimal bet sizes. Information abstraction and action abstraction are both explained in relation to their use in the game of poker.