Joint Post-Training Pruning and Power-of-Two Quantization for Efficient Edge Computing

Recent advancements in deep neural networks have created significant challenges for deploying these models on edge devices due to their computational and memory demands. We propose a novel integrated compression framework that combines nonlinear orthogonality-based channel pruning with progressive power-of-two (PoT) quantization to achieve efficient model compression for edge computing. Our framework first employs Radial Basis Function (RBF) kernel-based nonlinear orthogonality measurement to identify and remove redundant channels while preserving essential feature representations, then applies a layer-wise progressive power-of-two quantization scheme that enables efficient hardware implementation through bit-shift operations. Comprehensive experiments on CIFAR-10 and ImageNet demonstrate the effectiveness of our approach. On VGG16 with CIFAR-10, our method achieves 92.36% accuracy while reducing model size by 98.7% and computational complexity by 98.4%. On ResNet50 with ImageNet, we maintain 75.01% accuracy while achieving 95.93% model size reduction and 97.24% computational complexity reduction. Our framework significantly outperforms existing methods in terms of compression ratio and hardware efficiency while maintaining competitive accuracy. Recommended citation:

@INPROCEEDINGS{11228510, author={Yue, Zongcheng and Ma, Sean Longyu and Sham, Chiu-Wing and Fu, Chong}, booktitle={2025 International Joint Conference on Neural Networks (IJCNN)}, title={Joint Post-Training Pruning and Power-of-Two Quantization for Efficient Edge Computing}, year={2025}, volume={}, number={}, pages={1-8}, keywords={Quantization (signal);Accuracy;Image coding;Computational modeling;Neural network compression;Hardware;Model compression;Computational complexity;Edge computing;Residual neural networks;Edge Computing;Neural Network Compression;Channel Pruning;Power-of-Two Quantization}, doi={10.1109/IJCNN64981.2025.11228510}}