Semantic parsing aims at mapping natural language utterances into structured meaning representations. In this work, we propose a structure-aware neural architecture which decomposes the semantic parsing process into two stages. Given an input utterance, we ﬁrst generate a rough sketch of its meaning, where low-level information (such as vari-able names and arguments) is glossed over. Then, we ﬁll in missing details by taking into account the natural language input and the sketch itself. Experimental results on four datasets characteristic of different domains and meaning rep-resentations show that our approach consistently improves performance, achieving competitive results despite the use of relatively simple decoders.
Authors: Li Dong, Mirella Lapata (University of Edinburgh)