Energy-Efficient Implementation of Spiking Recurrent Cells on FPGA

2026-05-11

AI Summary

This work addresses the trade-off between energy efficiency and computational complexity in deploying biologically plausible spiking neural networks (SNNs) on FPGA hardware. The authors propose a lightweight SNN model based on Spiking Recurrent Cell (SRC) neurons, which preserves rich neural dynamics while significantly reducing hardware overhead through mathematical simplification, fixed-point approximation, and elimination of costly operations such as floating-point arithmetic and tanh/exp functions. A complete SNN architecture is implemented in VHDL, incorporating techniques including piecewise approximation, scaled quantization, and direct weight storage in lookup tables (LUTs) for efficient deployment on an Artix-7 FPGA. The system achieves 96.31% accuracy on MNIST with a per-sample inference latency of only 1.74 ms, and after compression, attains 92.89% accuracy at an energy cost of just 0.45 mJ per sample, effectively balancing performance, energy efficiency, and biological interpretability.

Abstract

Spiking Neural Networks (SNNs) can reduce energy consumption compared to conventional Artificial Neural Networks (ANNs) when spiking activity is sparse and the neuron model is hardware-friendly. However, biologically faithful models are often too costly to implement on FPGAs, whereas very simple models (e.g., IR/LIF) sacrifice part of the neuronal dynamics. In this work, we present an FPGA accelerator for an SNN using Spiking Recurrent Cell (SRC) neurons, providing a trade-off between biological plausibility and hardware cost. We propose a set of mathematical simplifications that remove costly unary operators (\textit{tanh}, \textit{exp}) and avoid floating-point arithmetic through scaling and piecewise-defined approximations. The complete network is implemented in VHDL and validated using spiking traces derived from the MNIST dataset. The weight matrices computed off-line are stored directly in LUT-registers without any adaptation. This demonstrates the robustness of SRC cells. Experiments were conducted on an Artix-7 XC7A200T clocked at 100 MHz. The reference implementation achieves 96.31\% accuracy with a 220-image spiking trace and a processing time of 1.7424 ms per digit. We then investigate accuracy/energy trade-offs by reducing the spiking trace length and quantizing synaptic weights down to 4 bits, achieving 93.32\% accuracy at 0.55 mJ per digit (55 images, 5-bit weights) and 92.89\% at 0.45 mJ (44 images, 4-bit weights). These results show that SRC-based SNNs can deliver competitive performance with reduced energy consumption, while preserving richer neuronal dynamics than standard LIF/IR models.

Problem

• Spiking Neural Networks
• FPGA
• Energy Efficiency
• Neuronal Dynamics
• Hardware Implementation

Innovation

• Spiking Recurrent Cell
• FPGA acceleration
• energy-efficient SNN
• piecewise approximation
• fixed-point arithmetic