Hands-On Machine Learning on Google Cloud Platform

Recursive neural networks

A recursive network is just a generalization of a recurrent network. In a recurrent network, the weights are shared and dimensionality remains constant along the length of the sequence. In a recursive network, the weights are shared and dimensionality remains constant but at every node. The following figure shows what a recursive neural network looks like:

Recursive neural networks can be used for learning tree-like structures. They are highly useful for parsing natural scenes and language.