AI "Libratus" Beats Top Texas Hold’em Poker Pros - How It Works?

Researchers Noam Brown and Tuomas Sandholm of Carnegie Mellon University have created an AI called libratus. Like other AI system AlphaGO, Libratus has also entered race with humans. Will see if the artificial brain can defeat humans in their own game.

> After Google AI, Its AI Libratus Turn to Beat Masters of Texas Hold'em:

Unlike Chess and Go, where Google AI has shown its supremacy. Libratus must prove its worth in the game Texas Hold'em, a popular form of Poker.

carnegie mellon university, artificial intelligence, ai, libratus, texas hold'em, poker, ai wins

carnegie mellon university, artificial intelligence, ai, libratus, texas hold'em, poker, ai wins

> Researchers Previously Tested AI Version “Lengpudashi”:

Held in January this year, the 20-day competition. Against four human poker professionals involved 120,000 hands of poker. All this to win the first prize of $200,000. The researchers once again tested AI poker skills a few months later during a poker competition held in China. Where they used a new AI version called “Lengpudashi”.

> AI's victory in Texas Hold'em Poker:

The aim why AI's victory in Texas Hold'em Poker is different is as the game is an example of the poor resolution of the info game. Here, not a single player knows all the elements of the game at any time during the game. Games like Chess and Go are perfect information games in which all related information is available in front of the players.

> For AI Systems, Its Challenging to Master Such Games:

The most exciting about “No Limit Texas Hold'em” is, it requires players to notice the fully random movements of other players. Also, make sure, you do not get tricked by the ability to bluff other players. For an Artificial Intelligence system, it is a challenging task to master such games. Which means taking small losses to get higher profits at the end.

> Inner Workings of Libratus:

Researchers reveal the inner workings of their Artificial Intelligence. By playing poker in a new Research article published in the journal Science.

> Methods Used By AI Libratus:

The methods used by Libratus to master the game are not based on expert human knowledge. In fact, they are not designed explicitly for poker. These can be used for many faulty information games.