AI Show: What are the different types of machine learning?

Scott: Welcome to the AI show. What are those big questions? Usually people think about three different types, like reinforcement learning, unsupervised, or supervised. Susan: Reinforcement learning is learning from a series of actions where you get a series of choices and rewards along away. So, a classic one that's been in the news is AlphaGo. A large chunk of reinforcement learning techniques are used in there, specifically Monte Carlo tree research techniques. Scott: So, people play the game Go. AlphaGo is a machine playing Go, being very good at it, and beating the world's top Go players. Susan: Another one that's really fun and has popped up the last couple of years is based off of Atari games.