On Robustness of Neural Semantic Parsers
This work addresses the vulnerability of neural semantic parsers to adversarial examples, which is a problem for developers deploying these systems in real-world applications.
This paper empirically studies the robustness of neural semantic parsers against adversarial attacks, where perturbed utterances maintain the original meaning. They propose a scalable methodology to construct robustness test sets and evaluate state-of-the-art parsers and the effect of data augmentation.
Semantic parsing maps natural language (NL) utterances into logical forms (LFs), which underpins many advanced NLP problems. Semantic parsers gain performance boosts with deep neural networks, but inherit vulnerabilities against adversarial examples. In this paper, we provide the empirical study on the robustness of semantic parsers in the presence of adversarial attacks. Formally, adversaries of semantic parsing are considered to be the perturbed utterance-LF pairs, whose utterances have exactly the same meanings as the original ones. A scalable methodology is proposed to construct robustness test sets based on existing benchmark corpora. Our results answered five research questions in measuring the sate-of-the-art parsers' performance on robustness test sets, and evaluating the effect of data augmentation.