Researchers Introduce A New Algorithm For Faster Reinforcement Learning

Any effective data-driven method for deep reinforcement learning should be able to use data to pre-train offline while improving with online fine-tuning.

Link to Full Article: Read Here