The fastest processor artificial intelligence in a world using light is a Greek invention, the result of five years of research by his Department of Informatics. AUTHORIZATION. The research project has already attracted the attention of the international scientific community.

In particular, as explained in the foundation’s announcement, “the processor is based on the original photonic neuron architecture, which allows algebraic operations to be performed at very high speeds, while at the same time using original neural network training techniques for optimal adaptation of algorithms. to the special characteristics of light. The advantages of the AUTh photonic processor are that it operates at a frequency of up to 50 GHz, i.e. 25 times faster than today’s GPUs, while at the same time reducing power consumption by 10 times.

“That is, greater speed and significant energy savings are achieved. This is one of the EU’s top priorities – especially after price increases throughout the year – and it proves that a very large proportion of European funds go to research programs to save energy. Consider its function ChatGPT, artificial intelligence application and a large linguistic model, it is very accurate, ”observes, referring to“ K ”, about Professor departments of informatics Anastasios Tefasand one of the two research supervisors of the study, together with Associate Professor Mr. Nikos Pleros.

The advantage of photonics technology in the field of information processing using neural networks is that it uses data in the form of light and photon integration technologies to implement deep neural networks that try to “mimic” the function of the human brain and human neurons in the field of computers, with the aim of faster and more efficient processing of information. The specific field of study today is a booming technology industry that began intensive research about five years ago, with the AUTH computer science department among the top three pioneers along with two of America’s leading universities, the Massachusetts Institute of Technology (MIT). ) and Princeton University.

AUTH’s pioneering research in this area was based on an interdisciplinary collaboration between KEDEK AUTH’s Wireless and Photonic Systems and Networks Research Group and the Computational Intelligence and Deep Learning Research Group of AUTH’s IT department. “Interdisciplinarity is one of the key elements of our success,” says Mr. Tefas.

Their research in photonic neural networks for artificial intelligence applications has received significant funding from a number of European and national research programs, as well as from US start-up Celestial AI based in Silicon Valley, California.

Research results have been published in international scientific journals such as Nature Communications, Journal of Lightwave Technology, IEEE Journal of Selected Topics on Quantum Electronics and SPIE Advanced Photonics, as well as in international press such as ScienceX, PIC Magazine, Compound Semiconductor, Nanowerk and etc. In March 2023, part of the results were presented in three scientific presentations at the world’s largest Fiber Optic Communications Conference (OFC) in San Diego, USA. The three principal investigators on the ads were AUTH Computer Science PhD PhDs Apostolos Tsakiridis, Georgios Jamouyannis, and Christos Pappas, under the supervision of AUTH Computer Science Department Professor Nikolaos Pleros. Three PhD students were selected by the International Conference Committee from seven finalists among 350 students from all over the world to receive the prestigious OFC Corning International Student Award. Thus, three of the seven finalists of the global competition became students of the Department of Informatics at AUTI.