An ‘AI for Poker’ EECS talk will be hosted by Libratus programmer Noam Brown on Thursday, March 7th, at the MIT Patil/Kiva seminar room.
Carnegie Mellon University PhD candidate and Facebook AI research scientist Noam Brown will conduct a special seminar on “AI for Imperfect-Information Game Settings” this Thursday starting at 4:00pm ET.
Brown (pictured, right) worked alongside Dr. Tuomas Sandholm to create Poker Bot Libratus, which defeated four Heads-Up No Limit Hold’em specialists in the January 2017 Brains vs. AI Poker Challenge held at The Rivers Casino in Pittsburgh.
The talk will form part of three seminars slated to take place over the next week which will discuss Artificial Intelligence for Poker, Go, and StarCraft.
Talks at MIT on AI for Poker, Go, and Starcraft all within a few days of each other (with the talk on AlphaGo and AlphaStar unfortunately overlapping):https://t.co/Ce240WhtcXhttps://t.co/pdQEkpJ6T4https://t.co/AuHaf8rckn https://t.co/OA9RJaCqpN
— Noam Brown (@polynoamial) March 5, 2019
The MIT Electrical Engineering & Computer Science (EECS) page states that Thursday’s talk on poker “will explain why past techniques intended for perfect-information multi-agent and imperfect-information single-agent settings break down both in theory and in practice in imperfect-information multi-agent settings, and the advances in Libratus and [Noam Brown’s] later work that overcome those challenges.”
DETAILS: MIT ‘AI for Imperfect-information Game Settings’ Special Seminar (MIT-EECS)
Libratus was recently hired by the Pentagon to assist the United States military. It “can tackle simulations that involve making decisions in a simulated physical space,” according to Dr. Sandholm via WIRED magazine.
Poker AI fans who will not be able to physically attend the discussion can visit Noam Brown’s YouTube channel after the fact to watch the lecture at their leisure. Below we have embedded a related talk hosted by the PhD candidate in late 2017, in which he reviewed AI for Poker — and how Libratus was able to defeat poker pros Jason Les, Donger Kim, Daniel McAulay, and Jimmy Chou.
AI for Poker Applications in Real Money Games
AI for Poker has become a hot topic in recent years across the real money online poker landscape as operators seek out AI products for player training. Microgaming Managing Director of Poker Alex Scott told Part Time Poker this February that “poker site-created bots could move at a pace that is geared toward each specific player,” and that “bots would be a great way to ensure that liquidity was always available — at any time of the day — in any game you wanted to play.”
However, Scott also reminded our readers that “the UK is a large enough market to sustain a wide variety of games with human players, so the benefit is less obvious – it’s the smaller markets where this idea has the most potential.” It is also worth noting that such technology would require regulatory approval in order to be rolled out commercially — a process that could take years to fully realize.
AI for Poker Evolution
WATCH: Claudico vs. Doug Polk (YouTube)
The evolution of AI for Poker, along with other popular games such as Jeopardy!, Go and StarCraft, has resulted in artificial intelligence progressing leaps and bounds compared to the best human players in each field.
In 2011, IBM computer WATSON bested past Jeopardy! champions Brad Rutter and Ken Jennings, proving that artificial intelligence is capable of performing at speeds that the best human players cannot.
AI for Poker in a Live Setting?
Although it is unknown at this time whether poker bots will eventually make their way over to “live” poker games on a commercial/promotional level, such a venue could set the stage for future 6-handed or 9-handed NLHE competitions — considered to be one of the final bastions in which the best players could potentially beat cutting-edge AI software.
SOCIAL MEDIA CONSIDERATION (TWITTER): @CarnegieMellon, @MIT, @polynoamial, @BotsAintLoyal
FOLLOW US ON TWITTER: @dhubermex, @benefactumgames, @WoernlePoker, @gonzo787, @PartTimePoker
FEATURED IMAGE CREDIT: Dr. Tuomas Sandholm & Noam Brown — Brains vs. AI Poker Challenge 2017