Q-SNNs: Quantized Spiking Neural Networks
Abstract: Brain-inspired Spiking Neural Networks (SNNs) leverage sparse spikes to represent information and process them in an asynchronous event-driven manner, offering an energy-efficient paradigm for the next generation of machine intelligence. However, the current focus within the SNN community prioritizes accuracy optimization through the development of large-scale models, limiting their viability in resource-constrained and low-power edge devices. To address this challenge, we introduce a lightweight and hardware-friendly Quantized SNN (Q-SNN) that applies quantization to both synaptic weights and membrane potentials. By significantly compressing these two key elements, the proposed Q-SNNs substantially reduce both memory usage and computational complexity. Moreover, to prevent the performance degradation caused by this compression, we present a new Weight-Spike Dual Regulation (WS-DR) method inspired by information entropy theory. Experimental evaluations on various datasets, including static and neuromorphic, demonstrate that our Q-SNNs outperform existing methods in terms of both model size and accuracy. These state-of-the-art results in efficiency and efficacy suggest that the proposed method can significantly improve edge intelligent computing.
Primary Subject Area: [Content] Vision and Language
Relevance To Conference: This paper proposes lightweight and hardware-friendly Quantized Spiking Neural Networks (Q-SNNs) to advance multimedia and multimodal processing on resource-constrained devices. This advancement can be described in two aspects. Firstly, by considering the quantization of both synaptic weights and membrane potentials, Q-SNNs significantly reduce the memory footprint and computational complexity, making them suitable for resource-constrained multimedia systems. Secondly, inspired by information entropy theory, we propose a novel Weight-Spike Dual Regulation (WS-DR) method to respectively adjust the weight and spike distributions, which increases the information quantities within Q-SNNs, leading to high-performance outcomes. In conclusion, Q-SNNs offer an energy-efficient and promising solution for deploying high-performance multimedia systems.
Submission Number: 2928
Loading