Quantized Convolutional Neural Networks Through the Lens of Partial Differential Equations
This work addresses the challenge of efficient CNN deployment for sensitive, resource-constrained applications like autonomous driving, though it appears incremental as it builds on existing quantization methods with novel analysis.
The paper tackled the problem of improving quantized convolutional neural networks (CNNs) for deployment on low-resource devices by using a PDE-based perspective, resulting in stable quantized networks that behave similarly to non-quantized counterparts with fewer parameters and sometimes improved accuracy.
Quantization of Convolutional Neural Networks (CNNs) is a common approach to ease the computational burden involved in the deployment of CNNs, especially on low-resource edge devices. However, fixed-point arithmetic is not natural to the type of computations involved in neural networks. In this work, we explore ways to improve quantized CNNs using PDE-based perspective and analysis. First, we harness the total variation (TV) approach to apply edge-aware smoothing to the feature maps throughout the network. This aims to reduce outliers in the distribution of values and promote piece-wise constant maps, which are more suitable for quantization. Secondly, we consider symmetric and stable variants of common CNNs for image classification, and Graph Convolutional Networks (GCNs) for graph node-classification. We demonstrate through several experiments that the property of forward stability preserves the action of a network under different quantization rates. As a result, stable quantized networks behave similarly to their non-quantized counterparts even though they rely on fewer parameters. We also find that at times, stability even aids in improving accuracy. These properties are of particular interest for sensitive, resource-constrained, low-power or real-time applications like autonomous driving.