“Shake-Shake Regularization of 3-Branch Residual Networks”, 2017-03-15 (; similar):
Reduce overfit by replacing, in a 3-branch ResNet, the standard summation of residual branches by a stochastic affine combination
The method introduced in this paper aims at helping computer vision practitioners faced with an overfit problem. The idea is to replace, in a 3-branch ResNet, the standard summation of residual branches by a stochastic affine combination. The largest tested model improves on the best single shot published result on CIFAR-10 by reaching 2.86% test error.
Code is available at GitHub.
[Keywords: Computer vision, Deep learning, Supervised Learning]