Multi-Adversarial Learning for Cross-Lingual Word Embeddings

TL;DR

This paper introduces a multi-adversarial learning approach for cross-lingual word embeddings, addressing limitations of previous GAN-based methods by modeling piece-wise linear mappings, significantly improving performance for distant languages.

Contribution

It proposes a novel multi-adversarial method that induces multiple mappings to better capture complex relationships between distant language embeddings.

Findings

Improved bilingual lexicon induction for distant languages.

Outperforms previous single-mapping GAN methods.

Effective in unsupervised cross-lingual embedding alignment.

Abstract

Generative adversarial networks (GANs) have succeeded in inducing cross-lingual word embeddings -- maps of matching words across languages -- without supervision. Despite these successes, GANs' performance for the difficult case of distant languages is still not satisfactory. These limitations have been explained by GANs' incorrect assumption that source and target embedding spaces are related by a single linear mapping and are approximately isomorphic. We assume instead that, especially across distant languages, the mapping is only piece-wise linear, and propose a multi-adversarial learning method. This novel method induces the seed cross-lingual dictionary through multiple mappings, each induced to fit the mapping for one subspace. Our experiments on unsupervised bilingual lexicon induction show that this method improves performance over previous single-mapping methods, especially for distant languages.