SPARQ: Spiking Early-Exit Neural Networks for Energy-Efficient Edge AI
This provides an energy-efficient solution for real-time AI at the edge, though it is incremental as it combines existing techniques like spiking networks, quantization, and early exits.
The paper tackled the computational overhead and lack of input-adaptive control in spiking neural networks for edge AI by introducing SPARQ, a framework integrating spiking computation, quantization, and early exits, which achieved up to 5.15% higher accuracy, over 330 times lower energy, and over 90% fewer synaptic operations compared to baselines.
Spiking neural networks (SNNs) offer inherent energy efficiency due to their event-driven computation model, making them promising for edge AI deployment. However, their practical adoption is limited by the computational overhead of deep architectures and the absence of input-adaptive control. This work presents SPARQ, a unified framework that integrates spiking computation, quantization-aware training, and reinforcement learning-guided early exits for efficient and adaptive inference. Evaluations across MLP, LeNet, and AlexNet architectures demonstrated that the proposed Quantised Dynamic SNNs (QDSNN) consistently outperform conventional SNNs and QSNNs, achieving up to 5.15% higher accuracy over QSNNs, over 330 times lower system energy compared to baseline SNNs, and over 90 percent fewer synaptic operations across different datasets. These results validate SPARQ as a hardware-friendly, energy-efficient solution for real-time AI at the edge.