Conversational Recommender System

Paper Link


3 Main Modules -

  1. NLU - alyzing each user utterance, keeping track of the user’s dialogue history and constantly updating the user’s intention.
  2. Dialog Management (DM) - decides which action to take given the current state.
  3. NLG - Generate response to the user.

The lines of research in this paper is an intersection of Dialogue Systems (DS), Recommender Systems, Faceted Search and Reinforcement Learning (RL).


CNN Classification Architecture

Belief Tracker

Recommender System

Reinforcement Learning

Rules for user to play with agent

Kaushik Rangadurai

Code. Learn. Explore

Share this post