As AI reshapes our world, Professor Chai Yang, Associate Dean (Research) of the Faculty of Science and Professor of the Department of Applied Physics at PolyU, is turning to nature for innovative solutions. His pioneering research in sensory AI is creating systems that are smarter, more energy-efficient, low-latency and memory-optimised.

Professor Chai’s pioneering research efforts have led to several innovations that overcome crucial barriers in power consumption, latency and memory within sensory AI systems. These innovations promise to transform technologies from mobile devices to IoT sensors and edge computing, impacting smart cities, autonomous vehicles and industrial automation.

For his exceptional contributions, he has been named a 2024 Falling Walls Winner in the Engineering & Technology category for “Breaking the Wall of Efficient Sensory AI Systems”. This prestigious award, reviewed by global experts from over 1,000 entries across 52 countries, recognises his pioneering work in hardware architectures and optimisation techniques for sensory AI.

Advancing sensory computing

Professor Chai’s work addresses fundamental challenges in AI systems. “The proliferation of data from ubiquitously distributed sensors leads to massive increases in sensory terminals. It is crucial to partially shift computation tasks to the sensory terminals,” he explained. His research defines near-sensor concepts and in-sensor computing paradigms based on the physical distance between sensory and computing units, dividing functions into low-level and high-level processing.

Drawing inspiration from natural bioinspired sensory systems, Professor Chai and his team have developed sensors that adapt to different light intensities, emulating and even surpassing the human retina’s adaptability. This approach reduces hardware complexity and improves machine vision systems, earning recognition as one of the Top Ten Hong Kong Innovation Technology News in 2022.

Inspired by flying insects, Professor Chai has also pioneered optoelectronic graded neurons for perceiving dynamic motion. This innovation efficiently encodes temporal information at sensory terminals, advancing machine vision systems with minimal hardware resources and potential applications in autonomous vehicles and surveillance systems.

These findings, published in high-impact journals like Nature Electronics and Nature Nanotechnology and highlighted in Nature, IEEE Spectrum and more, have garnered attention from research teams worldwide.

Building on these successes, Professor Chai envisions developing cutting-edge microelectronic and nanoelectronic devices with new functionalities. “We intend to create imaging technology capable of perceiving three-dimensional (3D) depth, four-dimensional (4D) spatial-temporal and multiple spectral information, utilising bioinspired mechanisms to reduce power consumption and latency,” he shared.

Learn more about Professor Chai’s research focus in the video: https://www.youtube.com/watch?v=Lk7Rga3kSoc