# A Full Non-Monotonic Transition System for Unrestricted Non-Projective Parsing

## TL;DR

This paper introduces a fully non-monotonic transition system for unrestricted non-projective dependency parsing, leveraging a novel dynamic oracle to improve parsing accuracy across multiple languages.

## Contribution

It presents the first non-monotonic transition system based on the Covington algorithm and develops new dynamic oracles for training, enhancing parsing performance.

## Key findings

- Non-monotonic system outperforms monotonic in most languages
- Dynamic oracle improves training efficiency and accuracy
- Applicable to diverse linguistic datasets

## Abstract

Restricted non-monotonicity has been shown beneficial for the projective arc-eager dependency parser in previous research, as posterior decisions can repair mistakes made in previous states due to the lack of information. In this paper, we propose a novel, fully non-monotonic transition system based on the non-projective Covington algorithm. As a non-monotonic system requires exploration of erroneous actions during the training process, we develop several non-monotonic variants of the recently defined dynamic oracle for the Covington parser, based on tight approximations of the loss. Experiments on datasets from the CoNLL-X and CoNLL-XI shared tasks show that a non-monotonic dynamic oracle outperforms the monotonic version in the majority of languages.