Cross-Lingual Word Embedding Refinement by $\ell_{1}$ Norm Optimisation

TL;DR

This paper introduces a robust post-processing method for cross-lingual word embeddings using $ ext{l}_1$ norm optimization, significantly improving their quality across multiple languages and tasks.

Contribution

It proposes a novel $ ext{l}_1$ norm-based refinement technique that is agnostic to the original training process of CLWEs, enhancing their robustness and performance.

Findings

Outperforms four state-of-the-art baselines in bilingual lexicon induction.

Improves cross-lingual transfer for natural language inference tasks.

Effective across ten diverse languages and various training corpora.

Abstract

Cross-Lingual Word Embeddings (CLWEs) encode words from two or more languages in a shared high-dimensional space in which vectors representing words with similar meaning (regardless of language) are closely located. Existing methods for building high-quality CLWEs learn mappings that minimise the $\ell_{2}$ norm loss function. However, this optimisation objective has been demonstrated to be sensitive to outliers. Based on the more robust Manhattan norm (aka. $\ell_{1}$ norm) goodness-of-fit criterion, this paper proposes a simple post-processing step to improve CLWEs. An advantage of this approach is that it is fully agnostic to the training process of the original CLWEs and can therefore be applied widely. Extensive experiments are performed involving ten diverse languages and embeddings trained on different corpora. Evaluation results based on bilingual lexicon induction and cross-lingual transfer for natural language inference tasks show that the $\ell_{1}$ refinement substantially outperforms four state-of-the-art baselines in both supervised and unsupervised settings. It is therefore recommended that this strategy be adopted as a standard for CLWE methods.