ML//RL//AlphaGo

DeepMind, 2016. Beat Lee Sedol at Go — the game everyone said AI couldn't play for decades. Monte Carlo tree search + deep neural nets for position evaluation + policy network for move selection. AlphaGo Zero then learned from scratch with zero human games, surpassing the original. The moment AI became visceral for the public.

DeepMind, 2016. Beat Lee Sedol at Go — the game everyone said AI couldn't play for decades. Monte Carlo tree search + deep neural nets for position evaluation + policy network for move selection. AlphaGo Zero then learned from scratch with zero human games, surpassing the original. The moment AI became visceral for the public.