Libratus AI Defeated Top Pros in 20 Days of Poker Play
Libratus, an artificial intelligence that defeated four top professional poker players in no-limit Texas Hold’em earlier this year, uses a three-pronged approach to master a game with more decision points than atoms in the universe, researchers at Carnegie Mellon University report.
In a paper published online today by the journal Science, Tuomas Sandholm, professor of computer science, and Noam Brown, a Ph.D. student in the Computer Science Department, detail how their AI achieved superhuman performance by breaking the game into computationally manageable parts and, based on its opponents’ game play, fix potential weaknesses in its strategy during the competition.
AI programs have defeated top humans in checkers, chess and Go — all challenging games, but ones in which both players know the exact state of the game at all times. Poker players, by contrast, contend with hidden information: what cards their opponents hold and whether an opponent is bluffing.
In a 20-day competition involving 120,000 hands at Rivers Casino in Pittsburgh this past January, Libratus became the first AI to defeat top human players at Head’s-Up, No-Limit Texas Hold’em — the primary benchmark and longstanding challenge problem for imperfect-information game-solving by AIs.
Libratus beat each of the players individually in the two-player game and collectively amassed more than $1.8 million in chips. Measured in milli-big blinds per hand (mbb/hand), a standard used by imperfect-information game AI researchers, Libratus decisively defeated the humans by 147 mmb/hand. In poker lingo, this is 14.7 big blinds per game.
“The techniques in Libratus do not use expert domain knowledge or human data and are not specific to poker,” Sandholm and Brown said in the paper. “Thus, they apply to a host of imperfect-information games.” Such hidden information is ubiquitous in real-world strategic interactions, they noted, including business negotiation, cybersecurity, finance, strategic pricing and military applications.
Libratus includes three main modules, the first of which computes an abstraction of the game that is smaller and easier to solve than by considering all 10161 (the number 1 followed by 161 zeroes) possible decision points in the game. It then creates its own detailed strategy for the early rounds of Texas Hold’em and a coarse strategy for the later rounds. This strategy is called the blueprint strategy.
One example of these abstractions in poker is grouping similar hands together and treating them identically.
“Intuitively, there is little difference between a king-high flush and a queen-high flush,” Brown said. “Treating those hands as identical reduces the complexity of the game and, thus, makes it computationally easier.” In the same vein, similar bet sizes also can be grouped together.
But in the final rounds of the game, a second module constructs a new, finer-grained abstraction based on the state of play. It also computes a strategy for this subgame in real-time that balances strategies across different subgames using the blueprint strategy for guidance — something that needs to be done to achieve safe subgame solving. During the January competition, Libratus performed this computation using the Pittsburgh Supercomputing Center‘s Bridges computer.
When an opponent makes a move that is not in the abstraction, the module computes a solution to this subgame that includes the opponent’s move. Sandholm and Brown call this nested subgame solving. DeepStack, an AI created by the University of Alberta to play Heads-Up, No-Limit Texas Hold’em, also includes a similar algorithm, called continual re-solving. DeepStack has yet to be tested against top professional players, however.
The third module is designed to improve the blueprint strategy as competition proceeds. Typically, Sandholm said, AIs use machine learning to find mistakes in the opponent’s strategy and exploit them. But that also opens the AI to exploitation if the opponent shifts strategy. Instead, Libratus’ self-improver module analyzes opponents’ bet sizes to detect potential holes in Libratus’ blueprint strategy. Libratus then adds these missing decision branches, computes strategies for them, and adds them to the blueprint.
In addition to beating the human pros, Libratus was evaluated against the best prior poker AIs. These included Baby Tartanian8, a bot developed by Sandholm and Brown that won the 2016 Annual Computer Poker Competition held in conjunction with the Association for the Advancement of Artificial Intelligence Annual Conference. Whereas Baby Tartanian8 beat the next two strongest AIs in the competition by 12 (plus/minus 10) mbb/hand and 24 (plus/minus 20) mbb/hand, Libratus bested Baby Tartanian8 by 63 (plus/minus 28) mbb/hand. DeepStack has not been tested against other AIs, the authors noted.
“The techniques that we developed are largely domain independent and can thus be applied to other strategic imperfect-information interactions, including nonrecreational applications,” Sandholm and Brown concluded. “Due to the ubiquity of hidden information in real-world strategic interactions, we believe the paradigm introduced in Libratus will be critical to the future growth and widespread application of AI.”
The Latest on: Libratus
- A New Frontier of AI and Deep Learning Capabilitieson March 31, 2020 at 5:00 pm
Some HPC systems are making huge strides in deep learning capabilities. Libratus, an AI powered by the Pittsburgh Supercomputing Center’s Bridges computer, recently took on four professional poker ...
- 3 Industries Adapting to Major AI Advances in 2020on March 23, 2020 at 9:06 am
Something that was enough to strike fear into the heart of even the most experienced of poker players was the news that the Libratus computer was able to win thousands of hands against human ...
- Beating humans at their own gameon March 12, 2020 at 2:22 pm
Carnegie Mellon University's Libratus AI system spent 20 days playing 120,000 hands of No Limit Texas Hold'em Poker against four poker professionals. The human pros were no match for the machine ...
- 35 Innovators Under 35on January 10, 2020 at 10:31 am
Brown’s creation, known as Libratus, is essentially three AI systems in one. The first developed a strategy for poker by playing against itself over trillions of hands during several months of ...
- AI Bot is the First to Play StarCraft II at the Grandmaster Levelon October 31, 2019 at 9:21 am
These researchers put Pluribus’ predecessor, the two-player Libratus AI, through this kind of test, but after this intense testing, “even the top professionals were not able to beat ...
- Game Over, Humanson December 23, 2018 at 4:25 pm
But twenty years later, it still came as a surprise that an AI called Libratus was able to outbluff and outperform us at a game that’s often thought of as the most human of all: poker.
- Stunning AI Breakthrough Takes Us One Step Closer to the Singularityon October 20, 2017 at 11:09 am
champions in 2011, the aforementioned defeat of Lee Sedol in 2016, and most recently, the defeat of four professional no-limit Texas hold’em poker players at the hands of Libratus, an AI ...
- Libratus Poker Win Paves Way for Microsoft Minecraft Challengeon March 20, 2017 at 2:04 pm
When Libratus beat four established poker pros, it not only shocked the community but paved the way for a new generation of artificial intelligence (AI) challenges. Although the pursuit of truly ...
- Czech ‘DeepStack’ v. CMU’s ‘Libratus’ Fight for Top AI Dog Status When It Comes to Beating Poker Pros at Hold’emon March 7, 2017 at 1:20 pm
DeepStack, a computer program developed by scientists from the University of Alberta in collaboration with researchers from both Czech Technical University and Prague’s Charles University ...
- Pros Lose to AI Bot at Texas Hold'emon March 1, 2017 at 6:08 pm
Named Libratus, the bot won after 120,000 hands of heads-up no-limit Texas Hold’em, taking a pot of $1.4 million virtual dollars. It’s got us dreaming of a future World Series of Poker showdown.
via Google News and Bing News