Meta Reinforcement Learning
Chelsea explains using meta reinforcement learning to teach students game programming, where each game is a different environment. The algorithm rewards policies that expose bugs, applicable not only in education but also in robotics. Plans to open source the auto grader for Breakout and expand to other games are in progress.In this clip
From this podcast
Gradient Dissent - A Machine Learning Podcast
Shaping the World of Robotics with Chelsea Finn
Related Questions
What is meta-learning for learning?
What is the main topic of the clip Learning Robotics, Meta Learning from the episode Chelsea Finn on Meta Learning & Model Based Reinforcement Learning?
What is the clip Learning Robotics, Meta Learning about from the episode Chelsea Finn on Meta Learning & Model Based Reinforcement Learning?