The work, carried out in collaboration with the University of Vigo, was co-led by PRL-UPV researchers José Capmany and Andrés Macho. The research has successfully developed and experimentally demonstrated the foundations of a new light-based information theory, termed Analog Photonic Information (API).
The research proposes a new way to process information using integrated photonics, meaning circuits that work with light instead of electricity. This technology "would allow certain complex calculations to be performed much faster and more efficiently than current electronic systems, especially in tasks related to artificial intelligence, scientific simulations, or medical diagnosis," the same sources indicated.
"Until now, mathematical models for computation were designed first, and then technology was adapted to them. We have reversed the process: we have created a mathematical model specifically designed to leverage the real capabilities of current and future photonic technology," explained José Capmany.
This new light-based language would allow for reduced processing times and energy consumption in operations that currently require enormous computing power. "This would have a direct impact on data centers, whose energy demand grows every year due to the rise of artificial intelligence and massive information processing," they stated.
Potential applications of the system include the development of new drugs through molecular simulations, the improvement of autonomous driving systems, advanced robotics, or medical image processing. In the latter case, it could help accelerate tests like CT scans, reducing the time needed to generate and process diagnostic images.
The technology could also be applied in fields such as astronomy or defense, where enormous volumes of data are handled requiring real-time responses.
One of the most "relevant" aspects of the work is that this new photonic computing would be more error-tolerant and require fewer additional resources to correct errors than other emerging technologies, such as quantum computing. "This would facilitate its scalability and future implementation in real systems," he stressed.
"If we are successful, we will have laid the foundation for designing a whole new generation of photonic chips that would coexist with current electronic ones and could transform the way we process information," stated Andrés Macho.
