K{\o}psala: Transition-Based Graph Parsing via Efficient Training and Effective Encoding

TL;DR

K{ exto}psala is a multilingual transition-based graph parser that uses efficient training and encoding techniques, demonstrating effectiveness in both Meaning Representation Parsing and Enhanced Universal Dependencies tasks.

Contribution

The paper introduces a unified pipeline for graph parsing that leverages off-the-shelf models and a transition-based parser with multilingual BERT, achieving competitive results.

Findings

Post-fix would have placed us 4th in ranking

Unified pipeline effective for multiple parsing tasks

Relies on tokenized forms and multilingual BERT for encoding

Abstract

We present K{\o}psala, the Copenhagen-Uppsala system for the Enhanced Universal Dependencies Shared Task at IWPT 2020. Our system is a pipeline consisting of off-the-shelf models for everything but enhanced graph parsing, and for the latter, a transition-based graph parser adapted from Che et al. (2019). We train a single enhanced parser model per language, using gold sentence splitting and tokenization for training, and rely only on tokenized surface forms and multilingual BERT for encoding. While a bug introduced just before submission resulted in a severe drop in precision, its post-submission fix would bring us to 4th place in the official ranking, according to average ELAS. Our parser demonstrates that a unified pipeline is effective for both Meaning Representation Parsing and Enhanced Universal Dependencies.