This repository contains code for Human-like informative conversations: Better acknowledgements using conditional mutual information.
The code is being cleaned up and will be coming soon. If you would like the access prior to it, email me at ashwinp@cs.stanford.edu.