Hypotheses ranking for robust domain classification and tracking in dialogue systems

Abstract

We present a novel application of hypothesis ranking (HR) for the task of domain detection in a multi-domain, multiturn dialog system. Alternate, domain dependent, semantic frames from a spoken language understanding (SLU) analysis are ranked using a gradient boosted decision trees (GBDT) ranker to determine the most likely domain. The ranker, trained using… (More)

Topics

5 Figures and Tables

Cite this paper

@inproceedings{Robichaud2014HypothesesRF, title={Hypotheses ranking for robust domain classification and tracking in dialogue systems}, author={Jean-Philippe Robichaud and Paul A. Crook and Puyang Xu and Omar Zia Khan and Ruhi Sarikaya}, booktitle={INTERSPEECH}, year={2014} }