Package crino :: Module network

Module network

The network module provides some ready-to-use neural network architectures, along with pretraining and supervised learning methods.

The currently implemented neural network architectures are :

AutoEncoder and its counterpart OutputAutoEncoder
MultiLayerPerceptron
DeepNeuralNetwork
InputOutputDeepArchitecture

See their respective documentations for more details about their use.

See Also: criterion, network

Classes

[hide private]

MultiLayerPerceptron
A MultiLayerPerceptron (MLP) is one classical form of artificial neural networks, whichs aims at predicting one or more output states given some particular inputs.

AutoEncoder
An AutoEncoder is a neural network whichs aims at encoding its inputs in a smaller representation space.

OutputAutoEncoder

PretrainedMLP
A PretrainedMLP is a specialization of the MLP, where the layers are pretrained, for input part on the training examples (x)orforoutputpartonthetraininglabels(:math:mathbf{y}) using a Stacked AutoEncoder strategy.

DeepNeuralNetwork
A DeepNeuralNetwork (DNN) is a specialization of the MLP, where the layers are pretrained on the training examples (x)usingaStacked strategy.

InputOutputDeepArchitecture
An InputOutputDeepArchitecture (IODA) is a specialization of the DNN, where the layers are divided into three categories : the input layers, the link layer and the output layers.