Introduced on day 38
Embedding
A dense, fixed-length vector that represents an item (word, document, airport, user) so that similar items lie close together under a chosen similarity.
In detail
Embeddings come from factorising co-occurrence statistics (LSA), from predicting context (word2vec, GloVe), from contrastive training of sentence encoders, or from a lookup table learned with a task loss (entity embeddings for categorical features). What "similar" means is set by the training objective and corpus, which is why an embedding must be evaluated on the task it will serve.