Shifting Mean Activation Towards Zero with Bipolar Activation Functions
arXiv:1709.04054
Abstract
We propose a simple extension to the ReLU-family of activation functions that allows them to shift the mean activation across a layer towards zero. Combined with proper weight initialization, this alleviates the need for normalization layers. We explore the training of deep vanilla recurrent neural networks (RNNs) with up to 144 layers, and show that bipolar activation functions help learning in this setting. On the Penn Treebank and Text8 language modeling tasks we obtain competitive results, improving on the best reported results for non-gated networks. In experiments with convolutional neural networks without batch normalization, we find that bipolar activations produce a faster drop in training error, and results in a lower test error on the CIFAR-10 classification task.
References in corpus (10)
- Adam: A Method for Stochastic Optimization
- Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
- Layer Normalization
- Deep Networks with Stochastic Depth
- All you need is a good init
- Zoneout: Regularizing RNNs by Randomly Preserving Hidden Activations
- Recurrent Dropout without Memory Loss
- Regularizing RNNs by Stabilizing Activations
- Path-Normalized Optimization of Recurrent Neural Networks with ReLU Activations
- Norm-preserving Orthogonal Permutation Linear Unit Activation Functions (OPLU)