Data-Driven Sparse Structure Selection for Deep Neural Networks

TL;DR

This paper introduces an end-to-end framework for learning and pruning deep neural networks by applying sparsity regularization to scaling factors, enabling automatic structure selection and model compression.

Contribution

It proposes a novel parameter-based sparsity regularization method with a modified APG optimizer for efficient network pruning without extensive trials.

Findings

Effective pruning of CNNs with minimal fine-tuning

Automatic adaptive depth and width selection

Comparable or superior performance to existing methods

Abstract

Deep convolutional neural networks have liberated its extraordinary power on various tasks. However, it is still very challenging to deploy state-of-the-art models into real-world applications due to their high computational complexity. How can we design a compact and effective network without massive experiments and expert knowledge? In this paper, we propose a simple and effective framework to learn and prune deep models in an end-to-end manner. In our framework, a new type of parameter -- scaling factor is first introduced to scale the outputs of specific structures, such as neurons, groups or residual blocks. Then we add sparsity regularizations on these factors, and solve this optimization problem by a modified stochastic Accelerated Proximal Gradient (APG) method. By forcing some of the factors to zero, we can safely remove the corresponding structures, thus prune the unimportant parts of a CNN. Comparing with other structure selection methods that may need thousands of trials or iterative fine-tuning, our method is trained fully end-to-end in one training pass without bells and whistles. We evaluate our method, Sparse Structure Selection with several state-of-the-art CNNs, and demonstrate very promising results with adaptive depth and width selection.