IJCAI 2020poster0 citations

Context Vectors Are Reflections of Word Vectors in Half the Dimensions (Extended Abstract)

Zhenisbek Assylbekov, Rustem Takhanov

Abstract

This paper takes a step towards the theoretical analysis of the relationship between word embeddings and context embeddings in models such as word2vec. We start from basic probabilistic assumptions on the nature of word vectors, context vectors, and text generation. These assumptions are supported either empirically or theoretically by the existing literature. Next, we show that under these assumptions the widely-used word-word PMI matrix is approximately a random symmetric Gaussian ensemble. This, in turn, implies that context vectors are reflections of word vectors in approximately half the dimensions. As a direct application of our result, we suggest a theoretically grounded way of tying weights in the SGNS model.

Natural Language Processing: EmbeddingsMachine Learning: Probabilistic Machine LearningMachine Learning: Tensor and Matrix MethodsMachine Learning: Unsupervised Learning
BibTeX
@inproceedings{ijcai2020p718,
  title     = {Context Vectors Are Reflections of Word Vectors in Half the Dimensions (Extended Abstract)},
  author    = {Assylbekov, Zhenisbek and Takhanov, Rustem},
  booktitle = {Proceedings of the Twenty-Ninth International Joint Conference on
               Artificial Intelligence, {IJCAI-20}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Christian Bessiere},
  pages     = {5115--5119},
  year      = {2020},
  month     = {7},
  note      = {Journal track},
  doi       = {10.24963/ijcai.2020/718},
  url       = {https://doi.org/10.24963/ijcai.2020/718},
}
Context Vectors Are Reflections of Word Vectors in Half the Dimensions (Extended Abstract) · IJCAI 2020