Statistical Dialogue Management using Intention Dependency Graph

We present a method of statistical dialogue management using a directed intention dependency graph (IDG) in a partially observable Markov decision process (POMDP) framework. The transition probabilities in this model involve information derived from a hierarchical graph of intentions. In this way, we combine the deterministic graph structure of a conventional rule-based system with a statistical dialogue framework. The IDG also provides a reasonable constraint on a user simulation model, which is used when learning a policy function in POMDP and dialogue evaluation. Thus, this method converts a conventional dialogue manager to a statistical dialogue manager that utilizes task domain knowledge without annotated dialogue data.