ShanghaiTech at MRP 2019: Sequence-to-Graph Transduction with Second-Order Edge Inference for Cross-Framework Meaning Representation Parsing

TL;DR

This paper introduces a graph-based parser for cross-framework meaning representation parsing, combining node generation with second-order edge inference, achieving top ranks in the CoNLL 2019 shared task.

Contribution

It proposes a novel sequence-to-graph transduction method integrating an extended pointer-generator and second-order inference for improved parsing accuracy.

Findings

Achieved 1st place in DM in-framework ranking

Achieved 2nd place in PSD in-framework ranking

Secured 3rd place in DM cross-framework ranking

Abstract

This paper presents the system used in our submission to the \textit{CoNLL 2019 shared task: Cross-Framework Meaning Representation Parsing}. Our system is a graph-based parser which combines an extended pointer-generator network that generates nodes and a second-order mean field variational inference module that predicts edges. Our system achieved \nth{1} and \nth{2} place for the DM and PSD frameworks respectively on the in-framework ranks and achieved \nth{3} place for the DM framework on the cross-framework ranks.