We propose a geometric framework for learning meta-embeddings of words from different embedding sources. Our framework transforms the embeddings into a common latent space, where, for example, simple averaging of different embeddings (of a given word) is more amenable. The proposed latent space arises from two particular geometric transformations - the orthogonal rotations and the Mahalanobis metric scaling. Empirical results on several word similarity and word analogy benchmarks illustrate the efficacy of the proposed framework.
@article{arxiv.2004.09219,
title = {Learning Geometric Word Meta-Embeddings},
author = {Pratik Jawanpuria and N T V Satya Dev and Anoop Kunchukuttan and Bamdev Mishra},
journal= {arXiv preprint arXiv:2004.09219},
year = {2020}
}