Filtered Batch Normalization
This work addresses a specific issue in neural network training for researchers and practitioners, offering an incremental improvement over standard batch normalization.
The paper tackles the problem of non-Gaussian activations in deep neural networks by proposing filtered batch normalization, which removes out-of-distribution activations to improve consistency in mean and variance calculations, resulting in faster convergence and higher validation accuracy.
It is a common assumption that the activation of different layers in neural networks follow Gaussian distribution. This distribution can be transformed using normalization techniques, such as batch-normalization, increasing convergence speed and improving accuracy. In this paper we would like to demonstrate, that activations do not necessarily follow Gaussian distribution in all layers. Neurons in deeper layers are more selective and specific which can result extremely large, out-of-distribution activations. We will demonstrate that one can create more consistent mean and variance values for batch normalization during training by filtering out these activations which can further improve convergence speed and yield higher validation accuracy.