ML 021: Grokking Deep Reinforcement Learning with Miguel Morales
Adventures in Machine Learning
00:00
How to Train a Dog, Right?
AI agents are not, you know, huge black boxes that we just put stuff in. Success stories such as AlphaGo cannot figure out like Starcraft is not able to be played. So the way to create a solution to a problem is first to actually model the problem. The same agent can train some problems better than others if it follows other conditions and assumptions for best learning.
Transcript
Play full episode
