Generative Models for Task Simulation
Minqi discusses the potential of using generative models as universal simulators for various tasks, emphasizing the idea of training open-ended agents within the design space captured by these models. The conversation also touches on the balance of prior knowledge in grounding exploration for open-ended learning.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#114 - Secrets of Deep Reinforcement Learning (Minqi Jiang)
Related Questions