Constructing Energy-efficient Mixed-precision Neural Networks through Principal Component Analysis for Edge Intelligence

Abstract

The `Internet of Things' has brought increased demand for AI-based edge computing in applications ranging from healthcare monitoring systems to autonomous vehicles. Quantization is a powerful tool to address the growing computational cost of such applications, and yields significant compression over full-precision networks. However, quantization can result in substantial loss of performance for complex image classification tasks. To address this, we propose a Principal Component Analysis (PCA) driven methodology to identify the important layers of a binary network, and design mixed-precision networks. The proposed Hybrid-Net achieves a more than 10% improvement in classification accuracy over binary networks such as XNOR-Net for ResNet and VGG architectures on CIFAR-100 and ImageNet datasets while still achieving up to 94% of the energy-efficiency of XNOR-Nets. This work furthers the feasibility of using highly compressed neural networks for energy-efficient neural computing in edge devices.

arXiv comment14 pages, 4 figures, 8 tables

References in corpus (12)

• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1
• DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
• Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
• PACT: Parameterized Clipping Activation for Quantized Neural Networks
• Learning both Weights and Connections for Efficient Neural Networks
• Mixed Precision Quantization of ConvNets via Differentiable Neural Architecture Search
• WRPN: Wide Reduced-Precision Networks
• Compression-aware Training of Deep Networks
• A Low Effort Approach to Structured CNN Design Using PCA
• Low-Precision Batch-Normalized Activations
• Per-Tensor Fixed-Point Quantization of the Back-Propagation Algorithm

Cited by in corpus (8)

• Exploring the Connection Between Binary and Spiking Neural Networks
• DIET-SNN: Direct Input Encoding With Leakage and Threshold Optimization in Deep Spiking Neural Networks
• AdderNet and its Minimalist Hardware Design for Energy-Efficient Artificial Intelligence
• QUANOS- Adversarial Noise Sensitivity Driven Hybrid Quantization of Neural Networks
• Conditionally Deep Hybrid Neural Networks Across Edge and Cloud
• Exposing the Robustness and Vulnerability of Hybrid 8T-6T SRAM Memory Architectures to Adversarial Attacks in Deep Neural Networks
• Complexity-aware Adaptive Training and Inference for Edge-Cloud Distributed AI Systems
• PBGen: Partial Binarization of Deconvolution-Based Generators for Edge Intelligence