Author: \/u\/NoamBrown<\/p>\n
Hi all! We are Noam Brown<\/a> and Professor Tuomas Sandholm<\/a>. We recently developed the poker AI Pluribus<\/a>, which has proven capable of defeating elite human professionals in six-player no-limit Texas hold’em poker, the most widely-played poker format in the world. Poker was a long-standing challenge problem for AI due to the importance of hidden information, and Pluribus is the first AI breakthrough on a major benchmark game that has more than two players or two teams. Pluribus was trained using the equivalent of less than $150 worth of compute and runs in real time on 2 CPUs. You can read our blog post on this result here<\/a>.<\/p>\n We are happy to answer your questions about Pluribus, the experiment, AI, imperfect-information games, Carnegie Mellon, Facebook AI Research, or any other questions you might have! A few of the pros Pluribus played against may also jump in if anyone has questions about what it’s like playing against the bot, participating in the experiment, or playing professional poker.<\/p>\n We are opening this thread to questions now and will be here starting at 10AM ET on Friday, July 19th to answer them.<\/p>\n<\/p><\/div>\n submitted by \/u\/NoamBrown <\/a> Go to Source<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"
[link]<\/a><\/span> [comments]<\/a><\/span><\/div>\n