Some Considerations on Learning to Explore via Meta-Reinforcement Learning

Add code
Mar 03, 2018
Bradly C. Stadie, Ge Yang, Rein Houthooft, Xi Chen, Yan Duan, Yuhuai Wu, Pieter Abbeel, Ilya Sutskever
Figure 1 for Some Considerations on Learning to Explore via Meta-Reinforcement Learning
Figure 2 for Some Considerations on Learning to Explore via Meta-Reinforcement Learning
Figure 3 for Some Considerations on Learning to Explore via Meta-Reinforcement Learning
Figure 4 for Some Considerations on Learning to Explore via Meta-Reinforcement Learning

Share this with someone who'll enjoy it:

Twitter IconFacebook IconLinkedin IconWhatsapp IconMessenger Icon

We consider the problem of exploration in meta reinforcement learning. Two new meta reinforcement learning algorithms are suggested: E-MAML and E-$\text{RL}^2$. Results are presented on a novel environment we call `Krazy World' and a set of maze environments. We show E-MAML and E-$\text{RL}^2$ deliver better performance on tasks where exploration is important.

View paper onarxiv iconopen_review iconOpenReview

Share this with someone who'll enjoy it:

Twitter IconFacebook IconLinkedin IconWhatsapp IconMessenger Icon