Efficient Deployment of Spiking Neural Networks on SpiNNaker2 for DVS Gesture Recognition Using Neuromorphic Intermediate Representation

2025-04-09

AI Summary

To address the insufficient co-optimization of accuracy and energy efficiency for spiking neural network (SNN) deployment on resource-constrained edge devices in DVS-based gesture recognition, this paper proposes a dual-path quantization method (PTQ/QAT) jointly optimizing synaptic weights and neuronal firing thresholds—marking the first incorporation of adaptive threshold scaling into quantization design, thereby departing from conventional weight-only quantization paradigms. Built upon the NIR framework, our approach employs 8-bit fixed-point arithmetic, percentile-based threshold scaling, and quantization-aware training, with deep architectural co-design for the SpiNNaker2 many-core neuromorphic platform. We establish the first SNN benchmark for DVS gesture recognition and achieve 94.13% on-chip accuracy on SpiNNaker2—nearly matching full-precision (32-bit floating-point) performance—while significantly reducing power consumption compared to state-of-the-art artificial neural network (ANN) solutions for the same task.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Hardware-aware ML
• Computer Vision: Learning & Optimization for CV

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments

Abstract

Spiking Neural Networks (SNNs) are highly energy-efficient during inference, making them particularly suitable for deployment on neuromorphic hardware. Their ability to process event-driven inputs, such as data from dynamic vision sensors (DVS), further enhances their applicability to edge computing tasks. However, the resource constraints of edge hardware necessitate techniques like weight quantization, which reduce the memory footprint of SNNs while preserving accuracy. Despite its importance, existing quantization methods typically focus on synaptic weights quantization without taking account of other critical parameters, such as scaling neuron firing thresholds. To address this limitation, we present the first benchmark for the DVS gesture recognition task using SNNs optimized for the many-core neuromorphic chip SpiNNaker2. Our study evaluates two quantization pipelines for fixed-point computations. The first approach employs post training quantization (PTQ) with percentile-based threshold scaling, while the second uses quantization aware training (QAT) with adaptive threshold scaling. Both methods achieve accurate 8-bit on-chip inference, closely approximating 32-bit floating-point performance. Additionally, our baseline SNNs perform competitively against previously reported results without specialized techniques. These models are deployed on SpiNNaker2 using the neuromorphic intermediate representation (NIR). Ultimately, we achieve 94.13% classification accuracy on-chip, demonstrating the SpiNNaker2's potential for efficient, low-energy neuromorphic computing.

Problem

• Optimize SNN deployment on SpiNNaker2 for DVS gesture recognition
• Address weight and neuron threshold quantization in SNNs
• Achieve efficient 8-bit inference with minimal accuracy loss

Innovation

• Uses SpiNNaker2 for SNN deployment
• Employs 8-bit quantization for efficiency
• Leverages NIR for model optimization