Learning Cross-lingual Embeddings from Twitter via Distant Supervision
This work addresses the challenge of adapting cross-lingual embeddings for social media applications, where informal and noisy text is common, offering a practical solution for tasks like multilingual sentiment analysis or content moderation.
The paper tackled the problem of learning cross-lingual embeddings from noisy social media text, specifically Twitter, and found that a simple post-processing step exploiting code-switching and shared vocabulary like emoji significantly improved the performance of state-of-the-art alignment methods, achieving gains of up to 5-10% in accuracy on benchmark tasks.
Cross-lingual embeddings represent the meaning of words from different languages in the same vector space. Recent work has shown that it is possible to construct such representations by aligning independently learned monolingual embedding spaces, and that accurate alignments can be obtained even without external bilingual data. In this paper we explore a research direction that has been surprisingly neglected in the literature: leveraging noisy user-generated text to learn cross-lingual embeddings particularly tailored towards social media applications. While the noisiness and informal nature of the social media genre poses additional challenges to cross-lingual embedding methods, we find that it also provides key opportunities due to the abundance of code-switching and the existence of a shared vocabulary of emoji and named entities. Our contribution consists of a very simple post-processing step that exploits these phenomena to significantly improve the performance of state-of-the-art alignment methods.