1. mateolopezareal

   

   Joined:

   Jun 4, 2020

   Posts:

   54

   Both of these params make the agent go random but what are the differences? Does Ml-agents use Epsilon-greedy Policy?

    

   Last edited: Jun 15, 2020

   mateolopezareal, Jun 15, 2020

   #1

2. vincentgao88

   

   Unity Technologies

   Joined:

   Feb 7, 2018

   Posts:

   21

   These are on totally different levels of training. To understand entropy you can refer to this article: https://medium.com/aureliantactics/ppo-hyperparameters-and-ranges-6fc2d29bccbe.
   To understand Curiosity you can refer to our blog post here: https://blogs.unity3d.com/2018/06/26/solving-sparse-reward-tasks-with-curiosity/

    

   vincentgao88, Jun 15, 2020

   #2

   mateolopezareal likes this.

Your name or email address:

Password:

Forgot your password?

Stay logged in