What is a word vector (how to calculate it), word embedding and how does word2vec work?

Question

Machine translation from one language to another

In relatively simple terms, please explain what a word vector is, how to calculate it. What is word embedding, skip-gramms, how does word2vec work? Preferably with examples.

Answer 1

a vector is a segment in space (a certain value of which we measure something in space).
If this is a 3-dimensional space, then this is, respectively, a set of parameters length, width, height of the object of which we measure.
The variety of words can be represented by a very large "space" (we can conditionally call it "meaning"), and WordToVek works by measuring the parameters of a certain word in this space.
For example, consider for example - if we limit the space of all words (meanings), only words with color - then each of our words is green, red, black, blue, etc., there will be a size in space that would correspond to the color characteristic (for example, RGB, such as there [0, 0, 0] - for black), however, our space is much larger than just color characteristics, such a space can have a thousand-dimensional dimension of "meanings".
the most important thing is that we eventually translate words into such vectors, we can compare them. find words that are close in meaning and compare the "figures" made up of these words in space

Answer 2

In fact, this is a whole technology. Expecting an explanation in "relatively simple words" is somehow not even serious. There is a lot of information on the topic - sit, read, understand, apply.
Here is a good - one of many - introduction to the topic:
https://skymind.ai/wiki/word2vec
with examples and explanations of "what is" and "how it works". If something specific becomes incomprehensible - ask.

Answer 3

so what exactly is unclear?
Google immediately gives hints
https://towardsdatascience.com/introduction-to-wor...
https://becominghuman.ai/how-does-word2vecs-skip-g...