Supervised Deep Sparse Coding Networks

TL;DR

This paper introduces a deep sparse coding network (SCN) that uses cascaded sparse coding modules for high discriminative power and low-dimensional clustering, trained end-to-end for image classification.

Contribution

The paper presents a novel deep sparse coding architecture with end-to-end supervised training, achieving competitive results with fewer parameters.

Findings

Achieves 5.81% error on CIFAR-10

Achieves 19.93% error on CIFAR-100

Uses cascaded sparse coding modules for feature extraction

Abstract

In this paper, we describe the deep sparse coding network (SCN), a novel deep network that encodes intermediate representations with nonnegative sparse coding. The SCN is built upon a number of cascading bottleneck modules, where each module consists of two sparse coding layers with relatively wide and slim dictionaries that are specialized to produce high dimensional discriminative features and low dimensional representations for clustering, respectively. During training, both the dictionaries and regularization parameters are optimized with an end-to-end supervised learning algorithm based on multilevel optimization. Effectiveness of an SCN with seven bottleneck modules is verified on several popular benchmark datasets. Remarkably, with few parameters to learn, our SCN achieves 5.81% and 19.93% classification error rate on CIFAR-10 and CIFAR-100, respectively.