Proxy-Normalizing Activations to Match Batch Normalization while Removing Batch Dependence
arXiv:2106.03743
Abstract
We investigate the reasons for the performance degradation incurred with batch-independent normalization. We find that the prototypical techniques of layer normalization and instance normalization both induce the appearance of failure modes in the neural network's pre-activations: (i) layer normalization induces a collapse towards channel-wise constant functions; (ii) instance normalization induces a lack of variability in instance statistics, symptomatic of an alteration of the expressivity. To alleviate failure mode (i) without aggravating failure mode (ii), we introduce the technique "Proxy Normalization" that normalizes post-activations using a proxy distribution. When combined with layer normalization or group normalization, this batch-independent normalization emulates batch normalization's behavior and consistently matches or exceeds its performance.
NeurIPS 2021 camera-ready
References in corpus (15)
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- A Learned Representation For Artistic Style
- Layer Normalization
- Modulating early visual processing by language
- High-Performance Large-Scale Image Recognition Without Normalization
- An Investigation into Neural Net Optimization via Hessian Eigenvalue Density
- Comparison of Batch Normalization and Weight Normalization Algorithms for the Large-scale Image Classification
- Complexity of Linear Regions in Deep Networks
- Batch Normalization is a Cause of Adversarial Vulnerability
- Normalizing the Normalizers: Comparing and Extending Network Normalization Schemes
- Rethinking "Batch" in BatchNorm
- Making EfficientNet More Efficient: Exploring Batch-Independent Normalization, Group Convolutions and Reduced Resolution Training
- Mean Shift Rejection: Training Deep Neural Networks Without Minibatch Statistics or Normalization
- Is Batch Norm unique? An empirical investigation and prescription to emulate the best properties of common normalizers without batch dependence
- Accelerating Training of Deep Neural Networks with a Standardization Loss