Transformer Semantic Parsing

Abstract

In neural semantic parsing, sentences are mapped to meaning representations using encoder-decoder frameworks. In this paper, we propose to apply the Transformer architecture, instead of recurrent neural networks, to this task. Experiments in two data sets from different domains and with different levels of difficulty show that our model achieved better results than strong baselines in certain settings and competitive results across all our experiments. We are thankful for our co-supervised student’s contribution. Namely, we express our gratitude to Xiang Li for his insight throughout his Bachelor of Advanced Computing (Honours) project (Li, 2019) in the Australian National University in 2019 that founded this study. We also thank the Australasian Language Technology Association and anonymous referees of its 2020 workshop for their helpful comments.