Deep-Unrolling Multidimensional Harmonic Retrieval Algorithms on Neuromorphic Hardware

TL;DR

This paper introduces a novel deep-unrolling algorithm for multidimensional harmonic retrieval, converting it into spiking neural networks for neuromorphic hardware, achieving significant power efficiency with minimal accuracy loss.

Contribution

It presents a new method to convert complex-valued CNNs for harmonic retrieval into energy-efficient SNNs on neuromorphic hardware, extending Few Spikes conversion for complex computations.

Findings

SNNs achieve nearly five times lower power consumption than CNNs.

The proposed method maintains moderate accuracy in harmonic retrieval tasks.

Conversion process effectively maps complex CNNs onto neuromorphic hardware.

Abstract

This paper explores the potential of conversion-based neuromorphic algorithms for highly accurate and energy-efficient single-snapshot multidimensional harmonic retrieval (MHR). By casting the MHR problem as a sparse recovery problem, we devise the currently proposed, deep-unrolling-based Structured Learned Iterative Shrinkage and Thresholding (S-LISTA) algorithm to solve it efficiently using complex-valued convolutional neural networks with complex-valued activations, which are trained using a supervised regression objective. Afterward, a novel method for converting the complex-valued convolutional layers and activations into spiking neural networks (SNNs) is developed. At the heart of this method lies the recently proposed Few Spikes (FS) conversion, which is extended by modifying the neuron model's parameters and internal dynamics to account for the inherent coupling between real and imaginary parts in complex-valued computations. Finally, the converted SNNs are mapped onto the SpiNNaker2 neuromorphic board, and a comparison in terms of estimation accuracy and power efficiency between the original CNNs deployed on an NVIDIA Jetson Xavier and the SNNs is being conducted. The measurement results show that the converted SNNs achieve almost five-fold power efficiency at moderate performance loss compared to the original CNNs.