MVR: Multi-view Video Reward Shaping for Reinforcement Learning
mvr-rl.github.io
MVR learns state relevance from multi-view video-text similarity and turns it into state-dependent reward shaping for reinforcement learning.
立即前往
返回首页
