Unsupervised Deep Learning by Injecting Low-Rank and Sparse Priors

TL;DR

This paper introduces an unsupervised deep learning approach that incorporates low-rank and sparse priors into neural networks, enabling effective learning without annotations, demonstrated through background subtraction with U-Net.

Contribution

It proposes a novel method to embed non-differentiable sparsity priors into deep networks using proximal mappings, facilitating unsupervised learning.

Findings

U-Net learns to detect moving objects without annotations.

The approach effectively captures high-dimensional data structure.

Successful background subtraction in test sequences.

Abstract

What if deep neural networks can learn from sparsity-inducing priors? When the networks are designed by combining layer modules (CNN, RNN, etc), engineers less exploit the inductive bias, i.e., existing well-known rules or prior knowledge, other than annotated training data sets. We focus on employing sparsity-inducing priors in deep learning to encourage the network to concisely capture the nature of high-dimensional data in an unsupervised way. In order to use non-differentiable sparsity-inducing norms as loss functions, we plug their proximal mappings into the automatic differentiation framework. We demonstrate unsupervised learning of U-Net for background subtraction using low-rank and sparse priors. The U-Net can learn moving objects in a training sequence without any annotation, and successfully detect the foreground objects in test sequences.