# Compressed Learning of Deep Neural Networks for OpenCL-Capable Embedded Systems

## TL;DR

This paper introduces a novel model compression framework for deep neural networks that enables efficient training and inference on embedded systems using OpenCL, leveraging sparse representations learned during training.

## Contribution

It presents a versatile compression method that does not require pre-trained models and improves effectiveness using proximal algorithms and debiasing techniques.

## Key findings

- Achieves minimal models suitable for embedded devices
- Uses sparse coding with proximal algorithms for better compression
- Does not require pre-trained models for compression

## Abstract

Deep neural networks (DNNs) have been quite successful in solving many complex learning problems. However, DNNs tend to have a large number of learning parameters, leading to a large memory and computation requirement. In this paper, we propose a model compression framework for efficient training and inference of deep neural networks on embedded systems. Our framework provides data structures and kernels for OpenCL-based parallel forward and backward computation in a compressed form. In particular, our method learns sparse representations of parameters using $\ell_1$-based sparse coding while training, storing them in compressed sparse matrices. Unlike the previous works, our method does not require a pre-trained model as an input and therefore can be more versatile for different application environments. Even though the use of $\ell_1$-based sparse coding for model compression is not new, we show that it can be far more effective than previously reported when we use proximal point algorithms and the technique of debiasing. Our experiments show that our method can produce minimal learning models suitable for small embedded devices.