NEUROPULS

Abstract

The Future of Computing: Unlocking the Potential of Neuromorphic Technology for Low-Power and Secure Edge Computing. With the rise of self-driving vehicles, IoT, and Industry 4.0, the need for processing massive amounts of data locally and securely has never been greater. Traditional electronic computing systems have limitations such as high latency and low energy-efficiency, which are no longer suitable for these applications. Neuromorphic computing, a brain-inspired approach, is the solution to these problems.

However, current neuromorphic electrical computing systems have their own limitations, which is why the NEUROPULS project is developing next-generation low-power and secure edge-computing systems. By utilizing novel photonic computing architectures and security layers based on photonic PUFs, augmented silicon photonics CMOS-compatible platforms, and emerging non-volatile phase change materials, this project aims to demonstrate a two-order of magnitude improvement in energy efficiency. With RISC-V compliant interfaces and a novel full-system simulation platform, NEUROPULS will revolutionize the future of computing.

Objectives

NEUROPULS will develop -for the first time- secure hardware accelerators based on novel neuromorphic architectures and PUF-based security layers leveraging the benefits offered by the integration of photonics, PCMs and III-V materials. This integration will provide superior security, energy-efficiency, and speeds for spiking and formal recurrent NNs when compared to current available technology for the selected use-cases.

Objective 1: Development of a CMOS-compatible platform addressing the integration of silicon photonics with PCMs and III-V materials
Objective 2: Development of a low-power and secure RISC-V interfaced neuromorphic accelerator based on the integration of silicon photonics, novel PCMs, and Q-switched III-V lasers
Objective 3: Development of a system-level simulation platform for PCM-based photonic low-power accelerators using photonic security layers