Learning a Wavelet-like Auto-Encoder to Accelerate Deep Neural Networks

TL;DR

This paper introduces a Wavelet-like Auto-Encoder that decomposes input images into low- and high-frequency channels, enabling neural network acceleration without structural modifications and preserving interpretability.

Contribution

The proposed WAE method allows neural network acceleration by decomposing inputs into frequency components and integrating with existing models, avoiding structural changes.

Findings

Enables acceleration of deep neural networks without modifying their architecture.

Provides an interpretable decomposition of input images into frequency components.

Achieves effective classification using low-frequency channels with lightweight fusion.

Abstract

Accelerating deep neural networks (DNNs) has been attracting increasing attention as it can benefit a wide range of applications, e.g., enabling mobile systems with limited computing resources to own powerful visual recognition ability. A practical strategy to this goal usually relies on a two-stage process: operating on the trained DNNs (e.g., approximating the convolutional filters with tensor decomposition) and fine-tuning the amended network, leading to difficulty in balancing the trade-off between acceleration and maintaining recognition performance. In this work, aiming at a general and comprehensive way for neural network acceleration, we develop a Wavelet-like Auto-Encoder (WAE) that decomposes the original input image into two low-resolution channels (sub-images) and incorporate the WAE into the classification neural networks for joint training. The two decomposed channels, in particular, are encoded to carry the low-frequency information (e.g., image profiles) and high-frequency (e.g., image details or noises), respectively, and enable reconstructing the original input image through the decoding process. Then, we feed the low-frequency channel into a standard classification network such as VGG or ResNet and employ a very lightweight network to fuse with the high-frequency channel to obtain the classification result. Compared to existing DNN acceleration solutions, our framework has the following advantages: i) it is tolerant to any existing convolutional neural networks for classification without amending their structures; ii) the WAE provides an interpretable way to preserve the main components of the input image for classification.