Mixup on first-convolution-layer feature maps via Siamese/triplet architecture outperforms input-image mixup for CNN generalization.
Improved generalization by adding both auto- association and hidden-layer noise to neural-network-based-classifiers
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Mixup of Feature Maps in a Hidden Layer for Training of Convolutional Neural Network
Mixup on first-convolution-layer feature maps via Siamese/triplet architecture outperforms input-image mixup for CNN generalization.