How to interpret the embedding in Deep Learning by TensorFlow

In this issue, the editor will bring you about how TensorFlow interprets the embedding in deep learning. The article is rich in content and analyzes and narrates it from a professional point of view. I hope you can get something after reading this article.

A set of words, these belong to discrete non-numerical objects, the basic requirement of numerical calculation is numerical, so they need to be mapped to real vectors.

Embedding is the process of digitizing discrete objects.

Embedding vectors, google's open source word2vec model does this, and now a few lines of code in TensorFlow can do this by calling API:

Word_embeddings = tf.get_variable ("word_embeddings"

[vocabulary_size, embedding_size])

Embedded_word_ids = tf.nn.embedding_lookup (word_embeddings, word_ids)

The shape of embedded_word_ids [vocabulary_size, embedding_size]

Visual display mainly needs to reduce the dimension of high-dimensional vectors.

Embedding can be trained through many types of networks and has a variety of loss functions and data sets. For example, for a large sentence corpus, recurrent neural networks can be used to predict the next word based on the previous word, and two networks can be trained for multilingual translation.

This is how the TensorFlow shared by the editor interprets the embedding in deep learning. If you happen to have similar doubts, you might as well refer to the above analysis. If you want to know more about it, you are welcome to follow the industry information channel.