Recursive neural network

<p>A recursive neural network is a kind of <a href="/facts/Deep_learning/JLuwD3ea">deep neural network</a> created by applying the same set of weights <a href="/facts/Recursion/CxSKxJoW">recursively</a> over a structured input, to produce a <a href="/facts/Structured_prediction/hqe45TQD">structured prediction</a> over variable-size input structures, or a scalar prediction on it, by traversing a given structure in <a href="/facts/Topological_sorting/VY7gNtYA">topological order</a>. These networks were first introduced to learn <a href="/facts/Distributed_representation/6V1jMlkx">distributed representations</a> of structure (such as <a href="/facts/Mathematical_logic/SaI7Y3wN">logical terms</a>), but have been successful in multiple applications, for instance in learning sequence and tree structures in <a href="/facts/Natural_language_processing/1hjMKsSN">natural language processing</a> (mainly continuous representations of phrases and sentences based on <a href="/facts/Word_embedding/7uRcBPqo">word embeddings</a>). 
</p>

Recursive neural network open-in-new

Recursive neural network