One Semantic Parser to Parse Them All: Sequence to Sequence Multi-Task Learning on Semantic Parsing Datasets

TL;DR

This paper explores multi-task learning architectures to unify various semantic parsing datasets, achieving competitive accuracy and improved generalization while significantly reducing model complexity.

Contribution

It introduces a shared multi-task learning model for multiple semantic parsing datasets, demonstrating improved performance and efficiency over single-task models.

Findings

Shared MTL model achieves comparable or better accuracy.

MTL reduces model parameters by 68%.

MTL shows better compositional generalization.

Abstract

Semantic parsers map natural language utterances to meaning representations. The lack of a single standard for meaning representations led to the creation of a plethora of semantic parsing datasets. To unify different datasets and train a single model for them, we investigate the use of Multi-Task Learning (MTL) architectures. We experiment with five datasets (Geoquery, NLMaps, TOP, Overnight, AMR). We find that an MTL architecture that shares the entire network across datasets yields competitive or better parsing accuracies than the single-task baselines, while reducing the total number of parameters by 68%. We further provide evidence that MTL has also better compositional generalization than single-task models. We also present a comparison of task sampling methods and propose a competitive alternative to widespread proportional sampling strategies.