Zoek in de site...

IMM colloquium by prof. Alex Khajetoorians: 'Brain-like circuitry capable of self-adaption at the single atom level' (Lecture)

Date
Tuesday 25 February 2020Add to my calendar
Time
from 16:00
Location
HG00.307
Speaker
prof. Alex Khajetoorians (Scanning Probe Microscopy)
Preceding lecture
Maarten van Eerden, MSc (Applied Materials Science): 'Observation of the Franz-Keldysh Effect in Ultra-Thin GaAs Solar Cells'
Description

Alex Khajetoorians

Information technology is rapidly leading to its own carbon footprint on our society. One of the culprits is the application of big data in deep learning and artificial intelligence, which in themselves are extremely expensive from an energy point of view. From a materials standpoint, our modern computing architecture is built upon the von Neumann concept, driven by the elegance of silicon-based transistors as well as magnetic-based storage. Nevertheless, it is becoming increasingly important to create new paradigms in materials research harnessing quantum phenomena in high-quality materials, based on unconventional concepts like brain-inspired computing. In this talk, I will introduce the ideas behind brain-inspired computing in solid-state materials. Starting from an overview of the requirements needed, inspired by physics-related machine learning, I will focus specifically on the application of magnetism phenomena toward this end. I will specifically discuss approaches in which we create model systems, utilizing scanning probe experiments at the single atom level combined with atomic-scale fabrication. Afterward, I will discuss an exciting finding, in which we have controllably coupling magnetic dopants in a semiconducting van der Waals material and realized a recurrent neural network, which is capable of self-learning.