Abstract:Due to the ability to implement customized topology, FPGA is increasingly used to deploy SNNs in both embedded and high-performance applications. In this paper, we survey state-of-the-art SNN implementations and their applications on FPGA. We collect the recent widely-used spiking neuron models, network structures, and signal encoding formats, followed by the enumeration of related hardware design schemes for FPGA-based SNN implementations. Compared with the previous surveys, this manuscript enumerates the application instances that applied the above-mentioned technical schemes in recent research. Based on that, we discuss the actual acceleration potential of implementing SNN on FPGA. According to our above discussion, the upcoming trends are discussed in this paper and give a guideline for further advancement in related subjects.

What problem does this paper attempt to address?

The paper aims to address the problem of efficiently implementing Spiking Neural Networks (SNNs) on hardware, particularly using Field-Programmable Gate Arrays (FPGAs) to deploy SNNs. With the increasing complexity and computational demands of SNNs, how to reduce power consumption and hardware costs while maintaining high performance has become a research focus. The paper reviews the current state-of-the-art SNN implementations and their applications on FPGAs, explores the practical acceleration potential of SNNs on FPGAs, and points out future research trends and development directions. Specifically, the paper addresses the following key issues: 1. **Hardware Implementation of SNN Models**: The paper collects widely used spiking neuron models, network structures, and signal encoding formats, and lists related hardware design schemes that are suitable for SNN implementation on FPGAs. 2. **Technical Challenges and Solutions**: The paper analyzes the technical difficulties in implementing SNN accelerators on FPGAs, including accelerator architecture, optimization methods, and future solutions. 3. **Application Examples**: The paper lists specific examples from recent research that apply the aforementioned technical schemes, demonstrating the practical application effects of SNNs in different scenarios. 4. **Future Trends**: Based on the above discussions, the paper explores the future development trends of SNN implementation on FPGAs and provides guidance for further research in related fields. In summary, through comprehensive review and analysis, the paper aims to provide researchers and engineers with the latest knowledge and technical support on how to efficiently implement SNNs on FPGAs.

A Survey of Spiking Neural Network Accelerator on FPGA

Spike Trains Encoding Optimization for Spiking Neural Networks Implementation in FPGA

A Survey of Neuromorphic Computing Based on Spiking Neural Networks

To Spike or Not to Spike? A Quantitative Comparison of SNN and CNN FPGA Implementations

A Survey of FPGA-Based Neural Network Accelerator

[DL] A Survey of FPGA-based Neural Network Inference Accelerators

A Low Power and Low Latency FPGA-Based Spiking Neural Network Accelerator

A survey of FPGA-based accelerators for convolutional neural networks

An Efficient Spiking Neural Network Accelerator with Sparse Weight.

Enabling Efficient On-Edge Spiking Neural Network Acceleration with Highly Flexible FPGA Architectures

An Event-driven Spiking Neural Network Accelerator with On-chip Sparse Weight

FireFly: A High-Throughput Hardware Accelerator for Spiking Neural Networks with Efficient DSP and Memory Optimization

Spiking Neural Network Implementation on FPGA for Multiclass Classification

Hardware implementation of spiking neural networks on FPGA

Advancements in Algorithms and Neuromorphic Hardware for Spiking Neural Networks

FPGA Implementation of Simplified Spiking Neural Network

Accelerating CNN inference on FPGAs: A Survey

Spiker+: a framework for the generation of efficient Spiking Neural Networks FPGA accelerators for inference at the edge

SPAT: FPGA-based Sparsity-Optimized Spiking Neural Network Training Accelerator with Temporal Parallel Dataflow

A Reconfigurable FPGA-based Spiking Neural Network Accelerator

Algorithms for Fast Spiking Neural Network Simulation on FPGAs