Iterative Utterance Segmentation for Neural Semantic Parsing

TL;DR

This paper introduces an iterative segmentation framework that enhances neural semantic parsers' ability to handle complex utterances by decomposing them into simpler spans without requiring extra labeled data.

Contribution

It proposes a novel, training-efficient framework that iteratively segments utterances and composes partial meanings, improving parsing accuracy across multiple domains.

Findings

Significant accuracy improvements on Geo, Formulas, and ComplexWebQuestions datasets.

Framework enables better compositional generalization without extra labeled data.

Consistent performance gains across different neural semantic parser models.

Abstract

Neural semantic parsers usually fail to parse long and complex utterances into correct meaning representations, due to the lack of exploiting the principle of compositionality. To address this issue, we present a novel framework for boosting neural semantic parsers via iterative utterance segmentation. Given an input utterance, our framework iterates between two neural modules: a segmenter for segmenting a span from the utterance, and a parser for mapping the span into a partial meaning representation. Then, these intermediate parsing results are composed into the final meaning representation. One key advantage is that this framework does not require any handcraft templates or additional labeled data for utterance segmentation: we achieve this through proposing a novel training method, in which the parser provides pseudo supervision for the segmenter. Experiments on Geo, ComplexWebQuestions, and Formulas show that our framework can consistently improve performances of neural semantic parsers in different domains. On data splits that require compositional generalization, our framework brings significant accuracy gains: Geo 63.1 to 81.2, Formulas 59.7 to 72.7, ComplexWebQuestions 27.1 to 56.3.