Part of what makes Texas hold ‘em poker intriguing to watch is that it’s a game where some of the cards are visible to all – both the players in the game and any spectators who are watching. This type of poker is what’s called an imperfect information game – all the players know different pieces of information about the elements in play. This kind of thinking has implications for technology, Slate reports.
Researchers at the University of Alberta have developed artificial intelligence that does a very good job winning at Texas hold ‘Em poker. But winning at poker is just the start. Software that could win an imperfect information game has potential applications in real-world situations that also have imperfect information, such as transportation systems. It’s an example of gamification research finding applications in artificial intelligence.
Lead researcher Michael Bowling tells Slate that a game, poker in particular, works well for developing and testing artificial intelligence algorithms because it offers scenarios that are both complicated and controlled. That means a computer can develop a strategy to work through those scenarios through trial and error.
We may not realize it but technology every day that employs complex algorithms to make decisions. Elevator controls are one example. But Bowling says that solutions in Texas hold ‘em poker could solve even bigger problems that require decision making, such as airport security checkpoints. The research is also finding applications in health care. At the University of Alberta Hospitals, researchers are developing diabetes management software that can make recommendations on diabetes treatments. Like the Texas hold ‘em poker scenarios, diabetes treatment also presents imperfect information – each patient’s situations can be different and patients are not always compliant with doctor orders. What doctors need is help determining the proper orders to give – which is not all that the decision-making process in a complex game.
“And that’s what our poker programs have to do, they have to be robust to ‘what are the cards my opponent has, and how does my opponent play,’” Bowling says.
The gamification research still needs more work. But the work done so far shows promise in translating the problem-solving techniques of poker into artificial intelligence technology with real-world applications.
Image credit: Wikimedia
