Episode 19: Minqi Jiang, UCL, on environment and curriculum design for general RL agents
Generally Intelligent
00:00
Active Learning as a Curriculum?
Active learning as a curriculum works quite well for fine evening, large line and if you basely an uncertainty based active sampling criteria. It's allso deeply ed two intrinsic rewards, in the sense that, like, if you just think about the space of all possible task s, justo beu like one giant m d p. Then cricula is basically just like, pushing you to where the rewards are. And so it's just densifying the reward signal overtime.
Transcript
Play full episode
