Backgammon

The first neural network that applied to backgammon was developed by Jerry Tasaro at IBM Labs. After each game, it tweaks the coefficients of all these different terms to be more in line with the result of the game. And then you do it again and again and again, and this endless tweaking process goes on. It's sort of Stasis Point where it believes that it has now mastered the game as well as it can given the set of coefficients.

Play episode from 25:34

Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!

Get the app