Awakening Latent Grounding from Pretrained Language Models for Semantic Parsing

TL;DR

This paper introduces an erasing-then-awakening method that reveals and leverages latent grounding in pretrained language models, significantly improving semantic parsing performance, especially in text-to-SQL tasks.

Contribution

It proposes a novel approach to uncover and utilize latent grounding in PLMs, enhancing their applicability to semantic parsing without requiring explicit grounding labels.

Findings

Successfully awakens understandable latent grounding in PLMs

Improves text-to-SQL parsing accuracy by up to 9.8%

Demonstrates potential for enhancing downstream semantic tasks

Abstract

Recent years pretrained language models (PLMs) hit a success on several downstream tasks, showing their power on modeling language. To better understand and leverage what PLMs have learned, several techniques have emerged to explore syntactic structures entailed by PLMs. However, few efforts have been made to explore grounding capabilities of PLMs, which are also essential. In this paper, we highlight the ability of PLMs to discover which token should be grounded to which concept, if combined with our proposed erasing-then-awakening approach. Empirical studies on four datasets demonstrate that our approach can awaken latent grounding which is understandable to human experts, even if it is not exposed to such labels during training. More importantly, our approach shows great potential to benefit downstream semantic parsing models. Taking text-to-SQL as a case study, we successfully couple our approach with two off-the-shelf parsers, obtaining an absolute improvement of up to 9.8%.