B. Gullett, J. Mowry, J. S. Plank, L. Whatley, C. P, Rizzo and C. D. Schuman

July, 2025

ICONS: International Conference on Neuromorphic Systems

https://iconsneuromorphic.cc/Schedule-2025/

View Article

Abstract

Spiking neural networks (SNNs) are attractive algorithms that pose numerous potential advantages over traditional neural networks. One primary benefit of SNNs is they may be run extremely efficiently on event-driven and braininspired neuromorphic hardware. SNNs’ low-energy characteristics make them ideal candidates for embedded systems that require low size, weight, and power (SWaP). However, neuromorphic hardware’s novelty means it is limited in both production and capabilities. Few SNN developers have access to neuromorphic computers that they can wire into their target embedded systems. In this work, we present the TENNLab neuromorphic framework embedder, a tool that translates an SNN into a portable and dependency-free inference code library intended for embedded von Neumann processing systems such as microcontrollers. The framework embedder aims to increase the accessibility of embedded neuromorphic computing by providing a means for users to evaluate their SNNs in low-SWaP systems without the high development times and monetary costs associated with true neuromorphic hardware. We delve into the framework embedder’s motivation, usage, implementation, and potential use cases. Additionally, we provide results for three example applications to contextualize the framework embedder’s performance as similar to other works that manually port SNNs onto microcontrollers. We hope that developers either use the framework embedder or write similar frameworks to ease the SNN deployment process for neuromorphic engineers across the world.

Citation Information

Text

author     B. Gullett and J. Mowry and J. S. Plank and L. Whatley and C. P, Rizzo and C. D. Schuman
title      Generating Spiking Neural Network Code Libraries for Embedded Systems
booktitle  International Conference on Neuromorphic Computing Systems (ICONS)
publisher  ACM
month      July
year       2025

Bibtex

@INPROCEEDINGS{gmp:25:gsn,
    author = "B. Gullett and J. Mowry and J. S. Plank and L. Whatley and C. P, Rizzo and C. D. Schuman",
    title = "Generating Spiking Neural Network Code Libraries for Embedded Systems",
    booktitle = "International Conference on Neuromorphic Computing Systems (ICONS)",
    publisher = "ACM",
    month = "July",
    year = "2025"
}