universal approximation theorem
GPTKB entity
Statements (19)
Predicate | Object |
---|---|
gptkbp:instanceOf |
gptkb:mathematical_concept
|
gptkbp:appliesTo |
feedforward neural networks
|
gptkbp:describes |
capability of neural networks to approximate functions
|
gptkbp:doesNotGuarantee |
generalization
efficient learning practical network size |
gptkbp:field |
gptkb:machine_learning
gptkb:mathematics neural networks |
gptkbp:generalizes |
deep learning theory
|
https://www.w3.org/2000/01/rdf-schema#label |
universal approximation theorem
|
gptkbp:provenBy |
gptkb:George_Cybenko
|
gptkbp:relatedTo |
function approximation
activation function |
gptkbp:requires |
sufficiently large hidden layer
|
gptkbp:state |
a feedforward network with a single hidden layer can approximate any continuous function under mild assumptions
|
gptkbp:yearProved |
1989
|
gptkbp:bfsParent |
gptkb:no_free_lunch_theorem
|
gptkbp:bfsLayer |
5
|