Robust Semantic Parsing with Adversarial Learning for Domain Generalization

TL;DR

This paper explores using adversarial learning to improve the robustness and domain generalization of semantic parsing models across different datasets and styles.

Contribution

It introduces a domain classification adversarial training method for semantic parsing without explicit domain knowledge, enhancing out-of-domain performance.

Findings

Adversarial learning improves model generalization on in-domain data.

Models show increased robustness to lexical and stylistic variations.

Performance gains are observed on both French and English semantic parsing benchmarks.

Abstract

This paper addresses the issue of generalization for Semantic Parsing in an adversarial framework. Building models that are more robust to inter-document variability is crucial for the integration of Semantic Parsing technologies in real applications. The underlying question throughout this study is whether adversarial learning can be used to train models on a higher level of abstraction in order to increase their robustness to lexical and stylistic variations.We propose to perform Semantic Parsing with a domain classification adversarial task without explicit knowledge of the domain. The strategy is first evaluated on a French corpus of encyclopedic documents, annotated with FrameNet, in an information retrieval perspective, then on PropBank Semantic Role Labeling task on the CoNLL-2005 benchmark. We show that adversarial learning increases all models generalization capabilities both on in and out-of-domain data.