English

Shape of Elephant: Study of Macro Properties of Word Embeddings Spaces

Computation and Language 2021-06-15 v1 Machine Learning

Abstract

Pre-trained word representations became a key component in many NLP tasks. However, the global geometry of the word embeddings remains poorly understood. In this paper, we demonstrate that a typical word embeddings cloud is shaped as a high-dimensional simplex with interpretable vertices and propose a simple yet effective method for enumeration of these vertices. We show that the proposed method can detect and describe vertices of the simplex for GloVe and fasttext spaces.

Keywords

Cite

@article{arxiv.2106.06964,
  title  = {Shape of Elephant: Study of Macro Properties of Word Embeddings Spaces},
  author = {Alexey Tikhonov},
  journal= {arXiv preprint arXiv:2106.06964},
  year   = {2021}
}

Comments

3 pages, 2 figures, EEML-2021 poster