RIT Awarded $422K Grant For PIC Technology
Researchers at the Rochester Institute of Technology will
use photonic integrated circuit technology to improve the processing speed and
energy consumption of brain-inspired computing technique through a $422,733
grant from the National Science Foundation.
PIC: Hybrid Silicon Electronic Photonic Integrated
Neuromorphic Networks is a multiyear project to advance neuromorphic computing
using photonic circuits. Neuromorphic computing, sometimes referred to as
brain-inspired computing, is a subfield of artificial intelligence where the
physical neural network architecture and its complex processing mechanisms are
inspired by the learning mechanisms in the human brain. This type of
architecture is currently developed using electronic integrated circuits, and the
research team will be applying similar methods using photonic devices.
The neuromorphic system will leverage the advantages of both
electronics and photonics to achieve higher performance and speed for devices
as well as lower energy consumption. Photonic implementations of neural
networks offer an advantage because light can easily perform computational
tasks that are traditionally challenging to do in electronic-only
"Electronic-only hardware, such as CMOS "“ a widely used
type of semiconductor "“ is not suitable for high-bandwidth applications
critical to our modern information world," said Stefan Preble, co-leader of the
research team. "But the internet is powered by photonic technologies "“ lasers,
electro-optic modulators, and photodetectors "“ because of light's high
bandwidth, speed, and low-energy consumption. This project aims to realize
high-performance neural networks using light."
Preble will be joined by Dhireesha Kudithipudi, professor of
computer engineering and an expert in neuromorphic computing and artificial
intelligence applications. In order to construct the neural networks for
photonic chips, the team will build upon known capabilities of electronics to
overcome the challenges of establishing better memory and amplification. This
hybrid approach, where electronics and photonics would be integrated, enables
the investigation of, and solutions for, the broadest class of problems in the
evolution of improved photonic chips.