acm-header
Sign In

Communications of the ACM

ACM Opinion

Intro to Model-Free and Model-Based Reinforcement Learning


Daeyeol Lee, professor of neuroscience at Johns Hopkins School of Medicine, Baltimore, Maryland.

“Computationally, model-based reinforcement learning is a lot more elaborate [than model-free RL].” -Daeyeol Lee.

Credit: Daeyeol Lee

Nearly every book on reinforcement learning (RL) contains a chapter that explains the differences between model-free and model-based reinforcement learning. But seldom are the biological and evolutionary precedents discussed in books about RL algorithms for computers.

However, Daeyeol Lee's The Birth of Intelligence, a book that explores the evolution of intelligence, offers an interesting explanation of model-free and model-based RL. In an interview, Lee, a neuroscientist, discusses different modes of RL in humans and animals, AI and natural intelligence, and future directions of research.

From TechTalks
View Full Article


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account