LAGr: Labeling Aligned Graphs for Improving Systematic Generalization in Semantic Parsing

TL;DR

This paper introduces LAGr, a graph-based semantic parsing method that improves systematic generalization by directly producing meaning representations as aligned graphs, outperforming traditional seq2seq models on key benchmarks.

Contribution

LAGr is a novel graph-based semantic parser that leverages aligned graphs for better systematic generalization, with both strongly- and weakly-supervised variants.

Findings

LAGr significantly outperforms seq2seq models on COGS and CFQ benchmarks.

Both supervised and weakly-supervised LAGr achieve notable improvements.

The approach enhances the ability to handle compositional generalization in semantic parsing.

Abstract

Semantic parsing is the task of producing a structured meaning representation for natural language utterances or questions. Recent research has pointed out that the commonly-used sequence-to-sequence (seq2seq) semantic parsers struggle to generalize systematically, i.e. to handle examples that require recombining known knowledge in novel settings. In this work, we show that better systematic generalization can be achieved by producing the meaning representation (MR) directly as a graph and not as a sequence. To this end we propose LAGr, the Labeling Aligned Graphs algorithm that produces semantic parses by predicting node and edge labels for a complete multi-layer input-aligned graph. The strongly-supervised LAGr algorithm requires aligned graphs as inputs, whereas weakly-supervised LAGr infers alignments for originally unaligned target graphs using an approximate MAP inference procedure. On the COGS and CFQ compositional generalization benchmarks the strongly- and weakly- supervised LAGr algorithms achieve significant improvements upon the baseline seq2seq parsers.