Poincaré GloVe: Hyperbolic Word Embeddings
This work addresses the need for better unsupervised word embeddings that reveal hierarchical relationships in language, offering a novel approach with broad applications in natural language processing.
The paper tackled the problem of learning word embeddings that capture hierarchical structures by embedding words in hyperbolic spaces, resulting in unsupervised embeddings that outperform strong baselines on similarity, analogy, and hypernymy detection tasks, with new state-of-the-art accuracy on WBLESS classification.
Words are not created equal. In fact, they form an aristocratic graph with a latent hierarchical structure that the next generation of unsupervised learned word embeddings should reveal. In this paper, justified by the notion of delta-hyperbolicity or tree-likeliness of a space, we propose to embed words in a Cartesian product of hyperbolic spaces which we theoretically connect to the Gaussian word embeddings and their Fisher geometry. This connection allows us to introduce a novel principled hypernymy score for word embeddings. Moreover, we adapt the well-known Glove algorithm to learn unsupervised word embeddings in this type of Riemannian manifolds. We further explain how to solve the analogy task using the Riemannian parallel transport that generalizes vector arithmetics to this new type of geometry. Empirically, based on extensive experiments, we prove that our embeddings, trained unsupervised, are the first to simultaneously outperform strong and popular baselines on the tasks of similarity, analogy and hypernymy detection. In particular, for word hypernymy, we obtain new state-of-the-art on fully unsupervised WBLESS classification accuracy.