I'm excited to share our latest work “Representation Learning Enables Scalable Multitask Deep Reinforcement Learning”.
In this work, we revisit a fundamental question in reinforcement learning:
What if representation learning was the key ingredient behind scalable RL?
1/🧵
