Deep Learning Essentials

Example use cases

With the pre-trained Word2Vec embedding, the downstream applications can be many for NLP, for example, document classification or sentiment classification. One example is called Doc2Vec, which, in the simplest form, takes the Word2Vec vectors of every word in the document and aggregates them by either taking a normalized sum or arithmetic mean of the terms. The resulting vector for each document is used for text classification. This type of application can be thought of as the direct application of the learned word embeddings.

On the other hand, we can apply the idea of Word2Vec modeling to other applications, for example, finding similar items in an e-commerce environment. During each session window, an online user may ...