“Shake-Shake Regularization of 3-Branch Residual Networks”, Xavier Gastaldi2017-03-15 (; 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]