Unsupervised Learning Potential
Minqi delves into estimating learning potential through regret in unsupervised environment design, highlighting the gap between agent performance and optimal performance. He contrasts the use of regret with value prediction error in reinforcement learning, shedding light on measuring learning potential in revisiting levels.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#114 - Secrets of Deep Reinforcement Learning (Minqi Jiang)
Related Questions