TeachMyAgent: a Benchmark for Automatic Curriculum Learning in Deep RL
Clement Romac * 1 Remy Portelas * 1 Katja Hofmann 2 Pierre-Yves Oudeyer 1
Abstract agent is unlikely to be optimal for complex learning prob-
lems. Building upon this, curriculum learning has long been
Training autonomous agents able to generalize identified as a key component for many machine learning
to multiple tasks is a key target ...
附件列表