feedforward neural network

GPTKB entity

Predicate	Object
gptkbp:instanceOf	gptkb:convolutional_neural_network
gptkbp:contrastsWith	gptkb:convolutional_neural_network recurrent neural network
gptkbp:firstDescribed	gptkb:Frank_Rosenblatt
gptkbp:hasActivationFunction	gptkb:ReLU sigmoid tanh
gptkbp:hasApplication	gptkb:dictionary function approximation pattern recognition regression
gptkbp:hasArchitectureType	feedforward
gptkbp:hasConnectionType	unidirectional
gptkbp:hasLayerType	hidden layer input layer output layer
gptkbp:hasModel	gptkb:universal_approximation_theorem
gptkbp:hasNoCycles	true
gptkbp:hasType	gptkb:reinforcement_learning supervised learning unsupervised learning
https://www.w3.org/2000/01/rdf-schema#label	feedforward neural network
gptkbp:input	feature vector
gptkbp:limitation	cannot handle cycles cannot model temporal sequences vanishing gradient problem
gptkbp:output	class label predicted value
gptkbp:parameter	weights biases
gptkbp:synonym	gptkb:multilayer_perceptron_(MLP)
gptkbp:trainer	backpropagation gradient descent
gptkbp:usedIn	gptkb:machine_learning deep learning
gptkbp:bfsParent	gptkb:Neural_Probabilistic_Language_Model gptkb:artificial_neural_networks gptkb:convolutional_neural_network gptkb:DSSM
gptkbp:bfsLayer	5