Multitask Learning

Creating 3D Models from Flat 2D Images
March 10, 2020
Ambient Computing Expands
March 10, 2020

Multitask Learning

Pluribus learned to do several things at once and built its own strategy to win.

In the past year, researchers at Carnegie Mellon University and Facebook’s AI lab released a superhuman A.I. for multiplayer poker called Pluribus, and by the end of 10,000 hands against 12 professional players, it got very good at bluffing.

It marked a significant milestone in A.I. research.

Winning at games like checkers and chess entails computational challenges. In poker, there is an information challenge, since what’s needed to win is hidden from other players.

Poker also pits multiple players against each other and achieving victory is more complex than capturing game pieces. Pluribus learned to do several things at once and built its own strategy to win.