Public Research Seminar by Microelectronics Thrust - Novel Devices and Algorithms for Bio-inspired Computing

Brain-inspired algorithms such as artificial neural networks (ANN) have been deployed for a variety of artificial intelligence (AI) applications. These algorithms are often implemented on a conventional computer with a von Neumann architecture, which, unlike the brain, physically separates memory and processing units. This not only slows them down, but also raises their energy consumption significantly. Neuromorphic computing addresses these issues by drawing inspiration from the brain to design time- and energy-efficient hardware for data processing. In this talk, I will firstly briefly discuss the developments and challenges in neuromorphic computing based on emerging memory devices. I will then discuss my past and current work on the exploration of memory devices with new materials and bio-inspired computing algorithms: first, memristor and optical memory based on two-dimensional materials, second, physically reconstructing biological neuronal connectivity from neuronal recordings using memristor, last, spatial–temporal data processing with an optical memory-based reservoir computing.