Neural Semantic Parsing in Low-Resource Settings with Back-Translation and Meta-Learning

TL;DR

This paper introduces a novel neural semantic parsing method that leverages back-translation, rule-based initialization, and meta-learning to perform well in low-resource settings without extensive annotated data.

Contribution

It proposes a combined approach of rule-based initialization, back-translation, and meta-learning to improve semantic parsing with limited supervision.

Findings

Achieves competitive accuracy on WikiSQL without annotated programs

Effectively generalizes to examples beyond initial rules

Incrementally improves performance across multiple datasets

Abstract

Neural semantic parsing has achieved impressive results in recent years, yet its success relies on the availability of large amounts of supervised data. Our goal is to learn a neural semantic parser when only prior knowledge about a limited number of simple rules is available, without access to either annotated programs or execution results. Our approach is initialized by rules, and improved in a back-translation paradigm using generated question-program pairs from the semantic parser and the question generator. A phrase table with frequent mapping patterns is automatically derived, also updated as training progresses, to measure the quality of generated instances. We train the model with model-agnostic meta-learning to guarantee the accuracy and stability on examples covered by rules, and meanwhile acquire the versatility to generalize well on examples uncovered by rules. Results on three benchmark datasets with different domains and programs show that our approach incrementally improves the accuracy. On WikiSQL, our best model is comparable to the SOTA system learned from denotations.