Statements (25)
Predicate | Object |
---|---|
gptkbp:instanceOf |
activation function
|
gptkbp:advantage |
improves model fitting
reduces risk of dying ReLU problem |
gptkbp:aIsLearnable |
true
|
gptkbp:category |
nonlinear activation function
|
gptkbp:differentiable |
true
|
gptkbp:domain |
real numbers
|
gptkbp:form |
f(x) = x if x > 0 else a*x
|
gptkbp:fullName |
Parametric Rectified Linear Unit
|
gptkbp:generalizes |
gptkb:ReLU
|
https://www.w3.org/2000/01/rdf-schema#label |
PReLU
|
gptkbp:introduced |
gptkb:Kaiming_He
|
gptkbp:introducedIn |
2015
|
gptkbp:parameter |
learnable parameter
|
gptkbp:publishedIn |
Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
|
gptkbp:range |
real numbers
|
gptkbp:relatedTo |
gptkb:ELU
gptkb:ReLU gptkb:Leaky_ReLU |
gptkbp:usedFor |
neural networks
|
gptkbp:usedIn |
deep learning
|
gptkbp:bfsParent |
gptkb:Leaky_ReLU
gptkb:Parametric_ReLU gptkb:LeakyReLU |
gptkbp:bfsLayer |
7
|