Recurrent Neural NetworksIt is a representation learning method that can map words, sentences, paragraphs, and articles into the same vector space according to their corresponding semantics, that is, representing combinable information as meaningful vectors.

Recurrent Neural Network Classification

Recurrent neural networks include two types of artificial neural networks:

• Recurrent Neural Network: The connections between neurons form a directed graph;
• Recursive Neural Network: Use similar neural network structures recursively to construct more complex deep networks.

References:

【1】http://wiki.jikexueyuan.com/project/tensorflow-zh/tutorials/recurrent.html