Day: 18 April 2018

Uncategorised Loading...

Evolved Policy Gradients

We’re releasing an experimental metalearning approach called Evolved Policy Gradients, a method that evolves the loss function of learning agents, which can enable fast training on...

Weiterlesen