Image recognition based on lightweight convolutional neural network: Recent advances

Y Liu, J Xue, D Li, W Zhang, TK Chiew, Z Xu - Image and Vision Computing, 2024 - Elsevier
Image recognition is an important task in computer vision with broad applications. In recent
years, with the advent of deep learning, lightweight convolutional neural network (CNN) has …

Advancing spiking neural networks toward deep residual learning

Y Hu, L Deng, Y Wu, M Yao, G Li - IEEE transactions on neural …, 2024 - ieeexplore.ieee.org
Despite the rapid progress of neuromorphic computing, inadequate capacity and insufficient
representation power of spiking neural networks (SNNs) severely restrict their application …

Siman: Sign-to-magnitude network binarization

M Lin, R Ji, Z Xu, B Zhang, F Chao… - IEEE Transactions on …, 2022 - ieeexplore.ieee.org
Binary neural networks (BNNs) have attracted broad research interest due to their efficient
storage and computational ability. Nevertheless, a significant challenge of BNNs lies in …

Bag of tricks with quantized convolutional neural networks for image classification

J Hu, M Zeng, E Wu - arxiv preprint arxiv:2303.07080, 2023 - arxiv.org
Deep neural networks have been proven effective in a wide range of tasks. However, their
high computational and memory costs make them impractical to deploy on resource …

Neural network compression using binarization and few full-precision weights

FM Nardini, C Rulli, S Trani, R Venturini - arxiv preprint arxiv:2306.08960, 2023 - arxiv.org
Quantization and pruning are two effective Deep Neural Networks model compression
methods. In this paper, we propose Automatic Prune Binarization (APB), a novel …

Understanding weight-magnitude hyperparameters in training binary networks

J Quist, Y Li, J van Gemert - arxiv preprint arxiv:2303.02452, 2023 - arxiv.org
Binary Neural Networks (BNNs) are compact and efficient by using binary weights instead of
real-valued weights. Current BNNs use latent real-valued weights during training, where …