Statistical user simulation is an efficient and effective way to train and evaluate the performance of a (spoken) dialog system. In this paper, we design and evaluate a modular data-driven dialog simulator where we decouple the “intentional” component of the User Simulator from the Error Simulator representing different types of ASR/SLU noisy channel distortion. While the former is composed by a Dialog Act Model, a Concept Model and a User Model, the latter is centered around an Error Model. We test different Dialog Act Models and Error Models against a baseline dialog manager and compare results with real dialogs obtained using the same dialog manager. On the grounds of dialog act, task and concept accuracy, our results show that 1) datadriven Dialog Act Models achieve good accuracy with respect to real user behavior and 2) data-driven Error Models make task completion times and rates closer to real data.

Peer Reviewed