evhub comments on What are some non-purely-sampling ways to do deep RL?

[evhub]

This is really neat; thanks for the pointer!