This post will introduce the core concepts underlying various policy gradient algorithms. As opposed to previously introduced methods - which suffered from the curse of dimensionality - it is not necessary to analyze the full action space to update t...

Markdown supported · ⌘↵ to post
No responses yet.