Modeling ASR Ambiguity for Dialogue State Tracking Using Word Confusion Networks
Preprint
Abstract
Spoken dialogue systems typically use a list of top-N ASR hypotheses for inferring the semantic meaning and tracking the state of the dialogue. However ASR graphs, such as confusion networks (confnets), provide a compact representation of a richer hypothesis space than a top-N ASR list. In this paper, we study the benefits of using confusion networks with a state-of-the-art neural dialogue state tracker (DST). We encode the 2-dimensional confnet into a 1-dimensional sequence of embeddings using an attentional confusion network encoder which can be used with any DST system. Our confnet encoder is plugged into the state-of-the-art 'Global-locally Self-Attentive Dialogue State Tacker' (GLAD) model for DST and obtains significant improvements in both accuracy and inference time compared to using top-N ASR hypotheses.Keywords
All Related Versions
- Version 1, 2020-02-03, ArXiv
- Version 2, 2020-08-01, ArXiv
- Published version: .
This publication has 11 references indexed in Scilit:
- A Data Efficient End-to-End Spoken Language Understanding Architecture2020
- Schema-Guided Multi-Domain Dialogue State Tracking with Graph Attention Neural NetworksProceedings of the AAAI Conference on Artificial Intelligence, 2020
- Speech Model Pre-Training for End-to-End Spoken Language Understanding2019
- Curriculum-Based Transfer Learning for an Effective End-to-End Spoken Language Understanding and Domain Portability2019
- Neural Confnet Classification: Fully Neural Network Based Spoken Utterance Classification Using Word Confusion Networks2018
- Towards End-to-end Spoken Language Understanding2018
- Towards Universal Dialogue State Tracking2018
- Neural Belief Tracker: Data-Driven Dialogue State Tracking2017
- LatticeRnn: Recurrent Neural Networks Over Lattices2016
- Finding consensus in speech recognition: word error minimization and other applications of confusion networksComputer Speech & Language, 2000