Build a spiking neural network with resistance RAM.

Introduction

This project approaches creating a proposed neuromorphic computing method in Spiking Neural Networks (SNNs). SNNs are built upon biomimicry of spiking activity within the human brain and emulate it using memristor technology and modern machine learning theory. We aim to create novel, increasingly power-efficient approaches to AI inference. Developers will be split into hardware and software teams where the software team will approach training the model and modeling data to fit a spiking nature, and the hardware team will construct the required memristor array and functional “processor” used in the SNN.

We are looking for software developers with experience in neural network architectures, data manipulation, and hardware developers with experience in PCB Fabrication and the fundamentals of creating computer systems. Experience with SNNTorch or general knowledge of Memristors is a plus. If you have any questions or would like to reach out, feel free to email vraj.prajapati@mail.utoronto.ca or message @vpro on Discord.