Research output: Contribution to journal › Article
Comparison of Regularization Methods for ImageNet Classification with Deep Convolutional Neural Networks. / Smirnov, Evgeny A.; Timoshenko, Denis M.; Andrianov, Serge N.
In: AASRI Procedia, Vol. 6, 2014, p. 89-94.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Comparison of Regularization Methods for ImageNet Classification with Deep Convolutional Neural Networks
AU - Smirnov, Evgeny A.
AU - Timoshenko, Denis M.
AU - Andrianov, Serge N.
PY - 2014
Y1 - 2014
N2 - Large and Deep Convolutional Neural Networks achieve good results in image classification tasks, but they need methods to prevent overfitting. In this paper we compare performance of different regularization techniques on ImageNet Large Scale Visual Recognition Challenge 2013. We show empirically that Dropout works better than DropConnect on ImageNet dataset.
AB - Large and Deep Convolutional Neural Networks achieve good results in image classification tasks, but they need methods to prevent overfitting. In this paper we compare performance of different regularization techniques on ImageNet Large Scale Visual Recognition Challenge 2013. We show empirically that Dropout works better than DropConnect on ImageNet dataset.
KW - Deep Neural Networks
KW - Convolutional Neural Networks
KW - Dropout
KW - DropConnect
KW - ImageNet
M3 - Article
VL - 6
SP - 89
EP - 94
JO - AASRI Procedia
JF - AASRI Procedia
SN - 2212-6716
ER -
ID: 5699727