Learning to Map Context-Dependent Sentences to Executable Formal Queries

TL;DR

This paper introduces a context-aware model that maps conversational utterances to executable queries by leveraging interaction history and reference modeling, improving performance on flight planning data.

Contribution

The paper presents a novel model that combines interaction-level encoding and copying mechanisms to handle context-dependent query generation.

Findings

Model effectively incorporates interaction history.

Explicit reference modeling improves query accuracy.

Demonstrated benefits on ATIS flight planning interactions.

Abstract

We propose a context-dependent model to map utterances within an interaction to executable formal queries. To incorporate interaction history, the model maintains an interaction-level encoder that updates after each turn, and can copy sub-sequences of previously predicted queries during generation. Our approach combines implicit and explicit modeling of references between utterances. We evaluate our model on the ATIS flight planning interactions, and demonstrate the benefits of modeling context and explicit references.