Max-Pooling Dropout for Regularization of Convolutional Neural Networks
This work addresses regularization in deep learning for computer vision, but it appears incremental as it builds on existing dropout and pooling methods.
The paper tackles the unclear effect of dropout in pooling layers of convolutional neural networks by showing that max-pooling dropout is equivalent to random activation selection and proposing probabilistic weighted pooling for test-time model averaging, with empirical evidence validating its superiority.
Recently, dropout has seen increasing use in deep learning. For deep convolutional neural networks, dropout is known to work well in fully-connected layers. However, its effect in pooling layers is still not clear. This paper demonstrates that max-pooling dropout is equivalent to randomly picking activation based on a multinomial distribution at training time. In light of this insight, we advocate employing our proposed probabilistic weighted pooling, instead of commonly used max-pooling, to act as model averaging at test time. Empirical evidence validates the superiority of probabilistic weighted pooling. We also compare max-pooling dropout and stochastic pooling, both of which introduce stochasticity based on multinomial distributions at pooling stage.