Logo image
A Spiking Convolutional Neural Network Algorithm Based on Piecewise-Linear Complementary Leaky Integrate-and-Fire (PCLIF) Neurons and Hybrid Surrogate Gradient Training
會議論文

A Spiking Convolutional Neural Network Algorithm Based on Piecewise-Linear Complementary Leaky Integrate-and-Fire (PCLIF) Neurons and Hybrid Surrogate Gradient Training

Zong-Zhe Wu, Yi-Wen Chuang 和 Kea-Tiong Tang
Biomedical Circuits and Systems Conference, 頁碼.557-561
IEEE
2025 IEEE Biomedical Circuits and Systems Conference (BioCAS) (Abu Dhabi, United Arab Emirates, 16/10/2025–18/10/2025)
16/10/2025
Web of Science ID: WOS:001789406300118

摘要

Accuracy Convolutional neural networks Edge Computing Hybrid power systems Hybrid Training Neuromorphics Power demand SNN Neuron Spiking Neural Network Spiking neural networks Stability analysis Surrogate Gradient Technological innovation Training Neurons
Spiking Neural Networks (SNNs) offer energyefficient computation but still face challenges such as training instability, gradient vanishing, and inefficient hardware deployment. We propose a spiking convolutional neural network (SCNN) framework that integrates three key innovations. First, a Parameterized Sigmoid Gradient with Softplus (PSGS) enables adaptive gradient shaping to enhance early stability and late convergence. Second, a Hybrid Surrogate Gradient Training (HSGT) method dynamically switches between smooth and sharp surrogate functions over time. Third, we propose a PiecewiseLinear Complementary Leaky Integrate-and-Fire (PCLIF) neuron that employs piecewise linear activation and simplified memory updates to reduce inference cost. Our model achieves 97.57% accuracy on DVS128 Gesture and 80.40% on CIFAR10-DVS using only 10 time steps. Compared to CLIF, PCLIF reduces area and power by 44.8% and 70.9%, respectively, while maintaining competitive accuracy. These results demonstrate that our co-design of training strategy and neuron architecture enables training effectiveness and efficient deployment in low-power neuromorphic systems.

相關連結

指標

1 檢視次數

詳細資料

Logo image