CUHK Engineering professor wins the Optica Foundation 20th Anniversary Challenge
Professor Huang Chaoran, Assistant Professor from The Chinese University of Hong Kong (CUHK)’s Department of Electronic Engineering has won the Optica (formerly OSA) Foundation 20th Anniversary Challenge for her pioneering research on integrated photonic neuromorphic processor enabled intelligent, energy-efficient signal processing for next-generation communication systems, becoming one of ten winning projects globally.
Global internet traffic continues to grow at exponential rates and, to keep up with its scale, optical communications systems will need to provide 10 times more communication capacity, while the signal processors that serve as the foundation for these systems must reduce energy per bit tenfold. This need for both processing capability and energy efficiency creates challenges for today’s systems, but Professor Huang’s new research may have the answer: an integrated photonic neuromorphic processor.
Professor Huang’s design employs the strengths of the intrinsic properties of photonics, deep learning architectures and integrated photonic technologies to create a photonic processor that is anticipated to be 10 times more energy efficient and reduce processing latency by 1,000 times. In addition, employing silicon photonics as the basis for the photonic processor makes it more cost-effective and accessible for wider industry exploration.
Professor Huang said, “A photonic neural network is a hardware realisation of a deep learning model and will allow us to use photonic devices and shifts to process optical communication signals faster and with more power efficiency. We plan to build a photonic processor that can provide a sustainable processing speed and energy efficiency to tackle these challenges. This work can have an immediate and transformative impact on the important application of optical communication.”
Building on existing research, Professor Huang expects to create a small prototype that offers a proof-of-concept demonstration of the photonic processor within six months. She will identify a way to engineer a photonic neuron that can provide a speed of 50GHz. She will also incorporate high-speed integrated photonic circuits and electronic control circuits to increase bandwidth and efficiency.