To the surprise of many, an AI programme has beaten human expert competitors to 1st place in a number of exhibition poker matches. Totalling 360,000 hands, a series of poker matches were hosted in China over a five-day period, battling engineers, computer scientists and investors (Team Dragon) against an AI programme.
Led by a venture capitalist and World Series of Poker veteran Alan Du, Team Dragon were strongly beaten out of the £230,000 prize money by AI programme Lengpudashi.
About the Winner
The Lengpudashi programme (translated into ‘cold poker master’) is an updated version of the AI programme, Libratus. Housed in a super-computing centre in Pittsburgh, the programme famously won more than £1.2m worth of chips when it defeated 4 poker experts at the Rivers Casino in Pittsburgh over a 20-day tournament back in January.
The AI programme used advanced algorithms and its ability to learn from each hand to develop winning strategies. One of the crucial reasons these AI programmes are entered into these difficult challenges is to test their strategic decision-making skills and increase their abilities.
One of the most exciting things about Lengpudashi’s win is that poker is, what’s commonly referred to as, an “imperfect information game”. This basically means that all the pieces of the game are not visible, unlike chess for example. So, in order to win, the programme not only has to figure out complex strategies but also has to learn how to bluff when it’s dealt a weaker hand. This was an element that people had thought computers would never learn – until now.
What happened to the prize money?
Although the AI programme certainly knows how to win money, its skills have not been stretched to knowing how to spend it. The prize money won in the poker match will therefore be invested in a firm named Strategic Machine, founded by Noam Brown and Tuomas Sandholm.
What does this mean for us?
Because AI algorithms and programmes have now been developed in a way that can utilise information and output strategy in a variety of scenarios, they could have a wider use in the business world; think negotiations, finance medical treatment and even cyber security.
This story also highlights how important investments are in big data analysis for increasing the speed of development for AI; this in turn could benefit many businesses in the future.