We are driving deep into the technological era, making unimaginable progress along the way in the field of #Artificial Intelligence. What was once considered only a tool for budding sci-fi writers has now transformed into one of the most controversial topics in the field of technology. Most people fear it, some revere its potential, while others continue to explore its capabilities. A.I has been taking various forms across various research institutes and has been tested with different purposes. It’s most recent assignment involved beating professional #Poker players at their own game and it managed to do so quite convincingly.
When a program called DeepBlue first beat the world’s chess Champion Garry Kasparov in 1997, the world was stunned into submission, finally recognizing a computer’s capabilities. Exactly twenty years later, a more advanced program called DeepStack has managed to beat professional poker players at Texas hold ‘em, but this time the news isn’t as shocking or unexpected.
The team responsible for creating DeepStack hails from the University of Alberta’s Computer Poker Research Group. The historic match between man and machine was conducted in December 2016, and the team recently published the data collected during the match in the Science journal.
The big difference between the program necessary to beat players in games like chess and checkers, and the one needed to beat players at Texas hold ‘em has to be with information and intuition. In a game like Chess, all the information is out on the board for both players to absorb and deduce their next move. For a game like Texas hold ‘em however, partial information is held by each of the players, thereby requiring the program to make more ‘intuition-based’ decision.
In order to win a game of poker, DeepStack has to understand various situations presented during the game, instead of relying on complete information like previous A.I programs. It therefore had to treat each situation within a poker game as mini-poker games, in order to succeed in the game as a whole. By understanding and solving these mini-games within any given game, the program was able to come out on top at the end.
Applications for DeepStack
According to the researchers behind the brilliant new algorithm, the applications of creating a program that understands imperfect information is immense. Almost all real-world problems are based on imperfect and incomplete data, which is why decision making based on the available information becomes quite difficult for human beings.
Artificial information can therefore be used to make sense of the imperfect data present in the real world, and provide solutions that are beyond human capabilities. We might have just taken our next big step towards creating machines that can strengthen human civilization instead of destroying it like some people fear. #Texas Hold Em