Notation sway absorption:

What is embedding anyway?

  • To image a discrete set of symbols such as word or character into a vector space with addition and scalar multiplication.
  • Mathematically.
    • In mathematics, embedding (うめこみ, embedding, imbedding) is a monjection such that the structure between mathematical structures is preserved.

    • Wikipedia
    • But in this case, the original set is just a discrete set, so there’s no particular structure to preserve.
  • Therefore, the expression “embedding vector” may not be very appropriate, but it is quite common to call the layer responsible for this process “embedding layer” in the field related to natural language processing by neural net. It can’t be helped.

This page is auto-translated from /nishio/埋め込みベクトル using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.