Splitting EUD graphs into trees: A quick and clatty approach

TL;DR

This paper introduces a method for converting EUD graphs into trees using sequence labeling, aiming for efficiency and exploring new ideas within limited development time.

Contribution

The approach of splitting EUD graphs into trees based on linguistic criteria and combining them with sequence labeling is a novel technique for EUD parsing.

Findings

Relatively poor results, indicating room for improvement.

System demonstrates feasibility of tree-splitting approach.

Potential for future enhancements with system polishing.

Abstract

We present the system submission from the FASTPARSE team for the EUD Shared Task at IWPT 2021. We engaged in the task last year by focusing on efficiency. This year we have focused on experimenting with new ideas on a limited time budget. Our system is based on splitting the EUD graph into several trees, based on linguistic criteria. We predict these trees using a sequence-labelling parser and combine them into an EUD graph. The results were relatively poor, although not a total disaster and could probably be improved with some polishing of the system's rough edges.