
Episode 19: Minqi Jiang, UCL, on environment and curriculum design for general RL agents
Generally Intelligent
00:00
Is There a Danger to This?
There's a lot of works that basically view two player games as an open ended learning setting, if your task face is complex enough. But the problem it openedius is that your process could get stuck. Even in the excel work, we do see this. So for the axcel work, we had dislike web base demo that lets you visualize the agent and let you like actually ondio. It lets you generate or using the sliders for each of the variables in the environment. And the task is for this bi pedal walker to perform continuous control till i walk across the whole landscape.
Transcript
Play full episode