Researchers from The Hong Kong University of Science and Technology (HKUST) have developed the world’s first all-optical neural network for deep machine learning.
This brings artificial intelligence (AI) one step closer to matching human brains in tackling complex problems such as pattern recognition or risk management, and at much lower energy consumption at the speed of light.
For a long time, optical computation has been limited to linear multiplication. With only linear multiplication, a neural network cannot be used for deep machine learning which simulates human brain functions.
Deep machine learning in AI requires a multilayer neural network in which nonlinear activation functions are necessary components. However, in a conventional hybrid optical neural network, nonlinear activation functions – which simulate the way neurons response in human brains – are implemented electronically, thereby restricting both the speed and power.
This is now changing as a research team led two professors from HKUST’s Department of Physics, demonstrated the first-of-its-kind multilayer all-optical artificial neural network, making large-scale optical neural networks a step closer to reality.
In addressing the challenges, the team used cold atoms with electromagnetically induced transparency – which only require very weak laser power – to carry out nonlinear optical activation functions and developed a two-layer fully-connected all-optical neural network.
The network was tested for classifying the order and disorder phases of the Ising model in condensed matter physics, and the outcomes were proven as accurate as a well-trained computer-based neural network.
While the work is a proof-of-principle demonstration, it shows that a new generation of optical artificial intelligence – which functions much faster at much lower energy consumption, is possible.
Going forward, the team hopes to expand the scale of this approach to build a bigger all-optical neural network with more complex architectures for practical applications such as image recognition.
Enabling AI and machine learning research in HK
There is a definite demand for more academic exploration into AI similar to the aforementioned research. Moreover, artificial intelligence (AI) is poised to disrupt every industry in every walk of human life.
To promote the healthy development of AI, HKUST recently established the Center for AI Research (CAiRE) – the first such centre among local universities, to spearhead cross-disciplinary research in all scientific, technological, societal, business and educational aspects of AI.
AI technologies have evolved rapidly over the past decade and are becoming a major driver of innovation. With AI and data science being a strategic research focus at HKUST, the university can play a role not only in advancing the technology but also in addressing the challenges it brings on security, privacy and ethics.
CAiRE can further consolidate the strength of HKUST’s faculty and researchers to bring out the benefits of AI for the good of the community.
The university is well-positioned to address the issues of AI, with its well-rounded faculty teams, specialized in speech, natural language processing, computer vision, machine learning, data analytics, as well as public policy, governance, social impacts and humanities.
The centre will propose new AI research projects, pulling in faculty from the four schools and research centres at HKUST – including the Big Data Institute, Robotics Institute, Human Language Technology Center and HKUST Shenzhen Institute.
CAiRE will also collaborate with top universities and organizations from Hong Kong and worldwide – such as MIT and the World Economic Forum to transfer AI research discoveries made by HKUST.
The Center is also planning to create new courses centred around ethics education on top of the standard AI curriculum now offered to the HKUST engineering students.