Speaker's Brief Introduction： Alexander Lvovsky is an experimental physicist. He was born and raised in Moscow and did his undergraduate in Physics at the Moscow Institute of Physics and Technology. In 1993, he became a graduate student in Physics at Columbia University in New York City. His thesis research, conducted under the supervision of Dr. Sven R. Hartmann, was in the field of coherent optical transients in atomic gases. After completing his Ph. D. in 1998, he spent a year at the University of California, Berkeley as a postdoctoral fellow in the Department of Physics, and then five years at Universität Konstanz in Germany, first as an Alexander von Humboldt postdoctoral fellow, then as a research group leader in quantum-optical information technology. In 2004 he became Professor in the Department of Physics and Astronomy at the University of Calgary, and from autumn 2018, a professor at the University of Oxford, where he conducts wide-profile research on quantum and optical technology. Alexander founded Lumai, a spin-off enterprise developing ultra-fast, energy-efficient optical neural network technology and COMPOS, a nationwide outreach program for UK state school students, offering high-level physics and mathematics tutorials to senior secondary school students. Alexander is a past Canada Research Chair, a lifetime member of the American Physical Society and a Fellow of the Optical Society. His accolades include the International Quantum Communications award, commendation letter from the Prime Minister of Canada and the Emmy Noether research award of the German Science Foundation. His work has been featured by CBC, NBC, Wired, New Scientist, MIT Technology Review, the Guardian and even Daily Mail.
Abstract: Optical neural networks (ONNs) harness the fundamental properties of light to enable ultrafast, energy-efficient computation, surpassing the limitations of digital-electronic systems in tasks such as large-scale matrix multiplications. By exploiting interference, diffraction, and nonlinearity, ONNs can perform parallel processing of high-dimensional data, reducing latency and power consumption. Their further benefit, not explored extensively to date, is the ability to preserve and utilize quantum properties of light. We take advantage of these properties to develop and use ONNs in imaging and computer vision applications, as well as to solve discrete optimization problems. ONNs can enable microscopes of unprecedented resolution, computer vision systems that can see in almost complete darkness and quantum machines for finding the best routes in logistics, creating efficient schedules, designing new materials, or managing investment portfolios.