Episode 17: Andrew Lampinen, DeepMind, on symbolic behavior, mental time travel, and insights from psychology
Generally Intelligent
00:00
The Type of Explanations Is Supervised, and That's a Critical Point
The tasks are supervised in a way that really specifically targets the task structure. They sort of explain the generalizable, abstract structure of the world and how the current situation fits into that. We always instantiate these tasks in a reinforcement learning setting. And it's partly because i think that increases the challenge with learning.
Transcript
Play full episode
