Conversational Semantic Parsing for Dialog State Tracking
This addresses the problem of accurately estimating user goals in conversational AI, with incremental improvements for dialog systems.
The paper tackled dialog state tracking by reformulating it as a semantic parsing task over hierarchical representations, resulting in a 20% improvement over state-of-the-art flat approaches.
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.