4 ms·
This is why all the Alpha%wildcard% projects use self play, that along with a fitness function, or score maximizer generates their training data rather than hav
by nerpderp82 3y ago
This is why all the Alpha%wildcard% projects use self play, that along with a fitness function, or score maximizer generates their training data rather than have to show it records of games. You just let it play the game.
- reaperman 3y agoNot all of them. AlphaStar for StarCraft was largely supervised learning, to mimic top human players. Once it was sufficiently mimicking existing replays, then it switched to Reinforcement Learning. That's why you see it do "nonsensical" things like destroying the "unbuildable plates" at the bottom of its natural ramp. This 100% wouldn't happen if it had only learned through self-play.