Deep Blue and AlphaGo: Two Paradigms
This video presents the same text shown beside it, spoken and on screen. It adds nothing the text does not say.
State
Deep Blue beat the world chess champion with massive search plus human-written evaluation rules and no learning; AlphaGo beat the world Go champion with learned judgment plus search.
Show
1997, New York: 200 million positions per second. 2016, Seoul: networks trained on human games and self-play.
Watch for
Neither machine examined everything — both were selective, and what selected differed.