NLReLU, f(x) = ln(β max(0,x) + 1), gives 0.16% to 2.04% higher CNN accuracy than ReLU on MNIST and CIFAR-10, with per-network tuned β.
Deep sparse re ctifier neural networks ,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Natural-Logarithm-Rectified Activation Function in Convolutional Neural Networks
NLReLU, f(x) = ln(β max(0,x) + 1), gives 0.16% to 2.04% higher CNN accuracy than ReLU on MNIST and CIFAR-10, with per-network tuned β.