Using 'PHIL' to control wavelength light

An example of the control signals which carry the weighted input signals and probe signals

In this paper*, led by this department and published in Nature Communications, the authors demonstrate controlling one wavelength light by using another, by way of an all-optical neuromorphic computing system based on time division multiplexing, capable of processing input vectors exceeding 250,000 elements within a unified framework.

The method involves a combination of nano antennas on a photonic waveguide and standing waves. Spatial separation selectively heats at single wavelengths, which is used to mediate the phase of other wavelengths, which do not interact with the nanoantennas, by integrating 50GHz signals.

This unified framework constitutes a leap towards large-scale photonic computing that satisfies the dimensional requirements of AI workloads.

*'All-optical temporal integration mediated by subwavelength heat antennas'.