Conversational Semantic Parsing for Dialog State Tracking
Computation and Language
2021-05-17 v3
Abstract
We consider a new perspective on dialog state tracking (DST), the task of estimating a user's goal through the course of a dialog. By formulating DST as a semantic parsing task over hierarchical representations, we can incorporate semantic compositionality, cross-domain knowledge sharing and co-reference. We present TreeDST, a dataset of 27k conversations annotated with tree-structured dialog states and system acts. We describe an encoder-decoder framework for DST with hierarchical representations, which leads to 20% improvement over state-of-the-art DST approaches that operate on a flat meaning space of slot-value pairs.
Keywords
Cite
@article{arxiv.2010.12770,
title = {Conversational Semantic Parsing for Dialog State Tracking},
author = {Jianpeng Cheng and Devang Agrawal and Hector Martinez Alonso and Shruti Bhargava and Joris Driesen and Federico Flego and Shaona Ghosh and Dain Kaplan and Dimitri Kartsaklis and Lin Li and Dhivya Piraviperumal and Jason D Williams and Hong Yu and Diarmuid O Seaghdha and Anders Johannsen},
journal= {arXiv preprint arXiv:2010.12770},
year = {2021}
}
Comments
Publish as a conference paper at EMNLP 2020