We then demonstrate a lower-bound performance guarantee on policies regularized by the stationary state distribution. In practice, SRPO can be an add-on module to context-based algorithms in both online and offline RL settings.

Neural Information Processing Systems http://nips.cc/

Then we could generally binarize the reduced feature map before fed into the GUST.

Humans are able to effortlessly group the world in terms of abstract and reusable'building blocks',