Semantic Parsing with Semi-Supervised Sequential Autoencoders

TL;DR

This paper introduces a semi-supervised sequential autoencoder approach for semantic parsing, leveraging generative models to improve performance in low-resource domains by utilizing unpaired logical forms.

Contribution

It presents a novel semi-supervised method combining generative models with sequence transduction for semantic parsing, especially effective with limited labeled data.

Findings

Improved semantic parsing accuracy in low-resource settings

Effective use of synthetically generated logical forms

Demonstrated applicability across multiple domains

Abstract

We present a novel semi-supervised approach for sequence transduction and apply it to semantic parsing. The unsupervised component is based on a generative model in which latent sentences generate the unpaired logical forms. We apply this method to a number of semantic parsing tasks focusing on domains with limited access to labelled training data and extend those datasets with synthetically generated logical forms.