Regularity Normalization: Neuroscience-Inspired Unsupervised Attention across Neural Network Layers

Abstract

Inspired by the adaptation phenomenon of neuronal firing, we propose the regularity normalization (RN) as an unsupervised attention mechanism (UAM) which computes the statistical regularity in the implicit space of neural networks under the Minimum Description Length (MDL) principle. Treating the neural network optimization process as a partially observable model selection problem, the regularity normalization constrains the implicit space by a normalization factor, the universal code length. We compute this universal code incrementally across neural network layers and demonstrate the flexibility to include data priors such as top-down attention and other oracle information. Empirically, our approach outperforms existing normalization methods in tackling limited, imbalanced and non-stationary input distribution in image classification, classic control, procedurally-generated reinforcement learning, generative modeling, handwriting generation and question answering tasks with various neural network architectures. Lastly, the unsupervised attention mechanisms is a useful probing tool for neural networks by tracking the dependency and critical learning stages across layers and recurrent time steps of deep networks.

Accepted and to be published by Entropy. Codes at https://github.com/doerlbh/UnsupervisedAttentionMechanism

References in corpus (6)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
• MINE: Mutual Information Neural Estimation
• The Price of Fair PCA: One Extra Dimension
• Biologically Inspired Mechanisms for Adversarial Robustness
• A Story of Two Streams: Reinforcement Learning Models from Human Behavior and Neuropsychiatry

Cited by in corpus (1)

• Geometric and Topological Inference for Deep Representations of Complex Networks