CREOL Faculty Explores High-Speed Computing Using Chip-Scale Photonic Technology
Published in Science Advances, the study explores how nonlinear photonic chips could support faster, more efficient information processing for future computing systems.
Facebook
LinkedIn
As artificial intelligence and machine learning applications continue to grow, researchers are exploring new ways to process information faster and more efficiently. Traditional electronic computing systems face increasing demands for speed and energy efficiency, creating opportunities for technologies that use light instead of electricity to perform computations.
A new study co-authored by CREOL Assistant Professor Midya Parto explores an approach to high-speed computing using photonic integrated circuits and optical parametric oscillators. These light-based systems can process information through complex optical interactions, offering a potential alternative to conventional electronics for certain machine learning tasks.
The research focuses on a technique known as reservoir computing, a form of machine learning that uses the natural dynamics of a physical system to process information. By harnessing nonlinear interactions on a photonic chip based on thin-film lithium niobate (TFLN), this work investigates how these systems could perform computational tasks at ultra-high speeds.
The findings contribute to ongoing efforts to develop compact, high-performance photonic computing technologies, with potential applications in artificial intelligence, signal processing and next-generation computing systems.
The study, “High-speed reservoir computing using photonic integrated circuit optical parametric oscillators” which was published in the journal Science Advances was carried out in collaboration with Alireza Marandi’s group at Caltech and co-authors Gordon Li, Ryoto Sekine, Robert Gray, Luis Ledezma, James Williams and Arkadev Roy.
Read the full research paper to learn more about the findings.