Multi-Adversarial Learning for Cross-Lingual Word Embeddings
This work addresses the challenge of cross-lingual word embeddings for distant languages, which is important for natural language processing applications, but it is incremental as it builds on existing GAN-based methods.
The paper tackles the problem of inducing cross-lingual word embeddings for distant languages, where existing GAN-based methods underperform due to assuming a single linear mapping. It proposes a multi-adversarial learning method that uses multiple piece-wise linear mappings, resulting in improved performance in unsupervised bilingual lexicon induction, especially for distant languages.
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.