microsoft / TextWorld

​TextWorld is a sandbox learning environment for the training and evaluation of reinforcement learning (RL) agents on text-based games.
Other
1.23k stars 188 forks source link

Algorithm documentation #174

Closed gari-marcos closed 5 years ago

gari-marcos commented 5 years ago

Hi The algorithm used in the act function of the Neural Agent in Building a simple agent is not explained. Would it be possible to document it? Thank you in advance.

MarcCote commented 5 years ago

Duplicate of #158