An A.I. bot that destroyed some of the most accomplished poker players humanity had to offer at heads-up no limit Texas hold’em in 2017 has heard the call of duty and enlisted in the US military.
According to public records unearthed this week by Wired, the bot, “Libratus,” and its creators at Carnegie Melon University were granted a two-year contract in August of up to $10 million to support the US Army’s Defense Innovation Unit.
The exact nature of Libratus’ work at the Pentagon is unknown.
Rise of the Machine
The bot caused shockwaves around the poker community when it wiped the floor with human professional poker players Jason Les, Dong Kim, Jimmy Chou and Daniel McAulay in a 180,000-hand heads-up challenge. Libratus battered its sentient opponents 11 hours a day for 20 days, winning a notional $1.8 million from them.
It was shocking because, while computers have been beating humans at chess since the 1980s, heads up no limit hold’em was supposed to be the game they would never master.
Unlike chess, it is a game of incomplete information with no “perfect strategy” and, as such, humans always been able to adapt their play more effectively than computers.
In a similar contest in 2015, Libratus’ predecessor, Claudico, came fifth out of five.
But Libratus used machine learning and computational game theory, playing trillions of poker hands against itself to compute an approximately perfect strategy.
“We didn’t tell Libratus how to play poker,” said Noam Brown of Carnegie Melon’s Computer Science Department. “We gave it the rules of poker and said, ‘learn on your own.’ The bot started playing randomly but over the course of playing trillions of hands was able to refine its approach and arrive at a winning strategy.
“When I see the bot bluff the humans, I’m like, ‘I didn’t tell it to do that. I had no idea it was even capable of doing that.’ It’s satisfying to know I created something that can do that,” he added.
But Libratus was never meant to be only a poker bot — it just so happened that mastering the heads-up no limit game was the perfect testing ground for its talents. It’s creators always said its future lay in other real-life applications where decisions have to be made using imperfect information. One such application was military strategy and planning, in war games and simulations.
In 2017, the Pentagon launched a program called Project Maven in response to then-US defense secretary James Mattis conviction that the military was falling behind the commercial sector in the use of A.I. and machine-learning technologies.
The same year Russian President Vladimir Putin declared A.I. to be the future “not only for Russia, but for all humankind.”
“It comes with colossal opportunities, but also threats that are difficult to predict. Whoever becomes the leader in this sphere will become the ruler of the world,” he said.