Skip to yearly menu bar Skip to main content


A Bregman Learning Framework for Sparse Neural Networks

Leon Bungert · Tim Roith · Daniel Tenbrinck · Martin Burger

Hall J (level 1) #1009

Keywords: [ JMLR ] [ Journal Track ]

Abstract: We propose a learning framework based on stochastic Bregman iterations, also known as mirror descent, to train sparse neural networks with an inverse scale space approach. We derive a baseline algorithm called LinBreg, an accelerated version using momentum, and AdaBreg, which is a Bregmanized generalization of the Adam algorithm. In contrast to established methods for sparse training the proposed family of algorithms constitutes a regrowth strategy for neural networks that is solely optimization-based without additional heuristics. Our Bregman learning framework starts the training with very few initial parameters, successively adding only significant ones to obtain a sparse and expressive network. The proposed approach is extremely easy and efficient, yet supported by the rich mathematical theory of inverse scale space methods. We derive a statistically profound sparse parameter initialization strategy and provide a rigorous stochastic convergence analysis of the loss decay and additional convergence proofs in the convex regime. Using only $3.4\%$ of the parameters of ResNet-18 we achieve $90.2\%$ test accuracy on CIFAR-10, compared to $93.6\%$ using the dense network. Our algorithm also unveils an autoencoder architecture for a denoising task. The proposed framework also has a huge potential for integrating sparse backpropagation and resource-friendly training. Code is available at

Chat is not available.