There have been rapid advancements in the field of Artificial Intelligence in the last few years, though this most recent development is astounding. Poker is largely a game of being able to read your opponents and get a sense for whether or not they’re bluffing. There’s definitely some statistics and math to the game, but Poker is a unique challenge for researchers as it forces the program to act using imperfect information.
“We need new AI techniques that can handle cases where decision-makers have different perspectives,” says Michael Bowling, professor in the University of Alberta’s Faculty of Science and principal investigator on the study. “Think of any real world problem. We all have a slightly different perspective of what’s going on, much like each player only knowing their own cards in a game of poker.”
The program is called Deepstack and it extends the ability to think about each situation during play using a technique called continual resolving. This allows Deepstack to make decisions based on the current situation rather than taking into account the entire game, a valuable skill when applied to the cards.
“We train our system to learn the value of situations,” says Bowling. “Each situation itself is a mini poker game. Instead of solving one big poker game, it solves millions of these little poker games, each one helping the system to refine its intuition of how the game of poker works. And this intuition is the fuel behind how DeepStack plays the full game.”
33 top Poker players were recruited and asked to play 3000 hands over the course of four weeks. 11 players finished the matches, and Deepstack beat every single one of them, making this the first time that a program has managed to beat professional in the field.
The ability for an Artificial Intelligence program to think on the fly and make judgments and adjustments is a breakthrough in the field, and may have ramifications in fields far beyond card games. Possible applications include making medical treatment recommendations, strategic defense planning, and even negotiation.