Implicit Compressibility of Overparametrized Neural Networks Trained with Heavy-Tailed SGD

Abstract

Neural network compression has been an increasingly important subject, not only due to its practical relevance, but also due to its theoretical implications, as there is an explicit connection between compressibility and generalization error. Recent studies have shown that the choice of the hyperparameters of stochastic gradient descent (SGD) can have an effect on the compressibility of the learned parameter vector. These results, however, rely on unverifiable assumptions and the resulting theory does not provide a practical guideline due to its implicitness. In this study, we propose a simple modification for SGD, such that the outputs of the algorithm will be provably compressible without making any nontrivial assumptions. We consider a one-hidden-layer neural network trained with SGD, and show that if we inject additive heavy-tailed noise to the iterates at each iteration, for any compression rate, there exists a level of overparametrization such that the output of the algorithm will be compressible with high probability. To achieve this result, we make two main technical contributions: (i) we prove a 'propagation of chaos' result for a class of heavy-tailed stochastic differential equations, and (ii) we derive error estimates for their Euler discretization. Our experiments suggest that the proposed approach not only achieves increased compressibility with various models and datasets, but also leads to robust test performance under pruning, even in more realistic architectures that lie beyond our theoretical setting.

31 pages, 2 figures

References in corpus (19)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
• A Mean Field View of the Landscape of Two-Layers Neural Networks
• On the Global Convergence of Gradient Descent for Over-parameterized Models using Optimal Transport
• Mean-field theory of two-layers neural networks: dimension-free bounds and kernel limit
• Compressible Distributions for High-dimensional Statistics
• Nonlinear SDEs driven by Lévy processes and related PDEs
• An Overview of Neural Network Compression
• Why are Adaptive Methods Good for Attention Models?
• Fractional Underdamped Langevin Dynamics: Retargeting SGD with Momentum under Heavy-Tailed Gradient Noise
• On the Heavy-Tailed Theory of Stochastic Gradient Descent for Deep Neural Networks
• Generalization bounds via distillation
• Quantitative Propagation of Chaos for SGD in Wide Neural Networks
• Heavy Tails in SGD and Compressibility of Overparametrized Neural Networks
• Compression based bound for non-compressed network: unified generalization error analysis of large compressible deep neural network
• Compressibility Loss for Neural Network Weights
• Neural Network Compression Via Sparse Optimization
• Deep neural networks with dependent weights: Gaussian Process mixture limit, heavy tails, sparsity and compressibility
• Compressing Heavy-Tailed Weight Matrices for Non-Vacuous Generalization Bounds