Artificial intelligence needs the speed of light

Posted date: 23 Oct 2019

The Physics department of UniTrento is working on all-optical devices for machine learning. The work is part of the Pelm project, whose objectives include the development of a cancer monitor. A non-invasive cancer monitor capable of monitoring the evolution of cancer tissues and therefore to provide information on the progress of the disease and the efficacy of the treatment.

That is one of the devices that will be developed within Pelm, the research project coordinated by the University of Trento that received almost 1 million euro funding over a period of three years by the Ministry of Education, universities and research as a Research project of national relevance (Prin). Coordinated by Lorenzo Pavesi of the Nanoscience laboratory of the Department of Physics of the University of Trento, the project uses light to build innovative all-optical platforms for artificial intelligence.

Machine learning processes, with which machines recognize images, understand human voice and carry out other basic tasks, have relied so far on microelectronic circuits and microprocessors stored on devices. These circuits are not efficient enough for artificial intelligence applications, use too much energy and this limits their use. The purpose of Pelm is to modify the notion of circuit to create energy-efficient, fast and non-invasive devices.

Posted By

GSP Admin