Implicit Word Reordering with Knowledge Distillation for Cross-Lingual Dependency Parsing

TL;DR

This paper introduces IWR-KD, a knowledge distillation approach that implicitly learns word reordering for cross-lingual dependency parsing, improving robustness across multiple languages without explicit reordering.

Contribution

It proposes a novel implicit reordering framework using knowledge distillation, addressing limitations of existing explicit reordering methods in cross-lingual parsing.

Findings

Outperforms existing methods on 31 languages

Enhances parser robustness to word order variations

Demonstrates effectiveness of implicit reordering with knowledge distillation

Abstract

Word order difference between source and target languages is a major obstacle to cross-lingual transfer, especially in the dependency parsing task. Current works are mostly based on order-agnostic models or word reordering to mitigate this problem. However, such methods either do not leverage grammatical information naturally contained in word order or are computationally expensive as the permutation space grows exponentially with the sentence length. Moreover, the reordered source sentence with an unnatural word order may be a form of noising that harms the model learning. To this end, we propose an Implicit Word Reordering framework with Knowledge Distillation (IWR-KD). This framework is inspired by that deep networks are good at learning feature linearization corresponding to meaningful data transformation, e.g. word reordering. To realize this idea, we introduce a knowledge distillation framework composed of a word-reordering teacher model and a dependency parsing student model. We verify our proposed method on Universal Dependency Treebanks across 31 different languages and show it outperforms a series of competitors, together with experimental analysis to illustrate how our method works towards training a robust parser.