So, Brown and Sandholm had to suffice with an approach to machine learning that is 'not guaranteed to converge to a Nash equilibrium.' It's a venture into the unknown in a sense, but one that wins nevertheless: 'ven though the techniques do not have known strong theoretical guarantees on performance outside of the two-player zero-sum setting, they are nevertheless capable of producing superhuman strategies in a wider class of strategic settings.'
As the authors write, 'Even approximating a Nash equilibrium is hard (except in special cases) in theory, and in games with more than two players, even the best complete algorithm can only address games with a handful of possible strategies per player.' Here's how it's related to artificial intelligence, how it works and why it matters.īut in multi-player Texas hold'em poker, the Nash Equilibrium becomes intractable computationally. What is machine learning? Everything you need to know
