Employing AI to streamline factory production

Artificial intelligence is revolutionising manufacturing. The technology is at the heart of the so-called Industrial Revolution 4.0, in which countless manufacturing processes are being entirely overhauled using automation and intelligent systems. It has long been a major area of focus for CUHK – and Dr. Shen Xiaoyong, a CUHK alumnus and entrepreneur, is one of its brightest stars.

Dr. Shen has long foreseen the development of AI technology. In 2019, he built a unicorn startup with Professor Jia Jiaya, his PhD supervisor and a world-class scientist specialising in computer vision and AI, which has developed an intelligent manufacturing inspection system. It can self-learn and adapt to different manufacturing environments, and manage tasks such as tracking materials, calculating and classifying workpieces, and performing quality control to increase efficiency. Their invention is paving the way for the smart factories of the future.

During his time at CUHK, Dr. Shen collaborated with a big tech firm to develop an AI-powered automatic object and background separation technology.

AI technology could revolutionise manufacturing

Dr. Shen was already aware of the potential of AI during his PhD studies at CUHK. In his sophomore year, in 2013, the frequent appearance of the term “Deep Learning” in hundreds of papers at an international conference drew his attention. It obviously was already a hot topic in academia. From that time onwards, he decided to commit himself to researching AI.   

“At that time, AI had just emerged and opportunities were everywhere. Professor Jia encouraged us to explore the relation between academic research and industrial application of the technology so that we could find a path of our own,” Dr. Shen recalled. “This is very important, as I came to realise the value of AI applications. Professor Jia and I talked a lot about the future development of AI technology and concluded that there was an enormous role for AI in the industrial sector.”　

Dr. Shen and Professor Jia Jiaya (right) built a unicorn startup focusing on the research and development of AI.

The majority of Dr. Shen’s research outputs are the result of collaborations with leading science and technology companies, laying the foundations for applying AI to industrial products and services. When he was a student, Dr. Shen worked with large enterprises such as Qualcomm, Huawei and Adobe. He contemplated the application of AI to industrial products, including the multi-lens fusion and optical zooming functions of mobile phone cameras, photo editing software, and the automatic object and background separation technology in online meeting software. After graduating in 2016, Dr. Shen joined Tencent’s X-lab with his mentor to develop a content recommendation system that suggests relevant videos to users based on their interests and preferences. They also worked on fully automatic makeup tools, autopilot software and intelligent analysis of medical images.    

The mentor-student duo starts a smart factory business

After two years of hard work and the experience they gathered working with the commercial sector, Dr. Shen and Professor Jia believed that industrial AI development had great potential. They hoped to develop solutions for manufacturers to help solve the pain points in product quality inspection.

The equipment developed by Dr. Shen and his company for semiconductor manufacturing companies to inspect chip quality.

In 2019, they left the company and set up their own business with several like-minded friends and CUHK PhD graduates. Dr. Shen explained that some semiconductor manufacturers were still using human vision inspections to monitor the quality of the chips, which was very labour intensive. “Workers have to inspect the chips under strong light, which is detrimental to their eyes; that is why they have to change shift every four hours. The problem, however, can be solved by AI technology, while demand for labour will also decrease accordingly. We have spent a lot of time studying the difficulties and the needs our clients have been encountering. Then we can work out the best solutions for them.”  

In only two years, their company has grown from a startup to a leading corporation in the AI industry with more than 600 employees, and frontier research and development centres and offices across Hong Kong, Shenzhen, Shanghai, Beijing, Suzhou, Hangzhou, Chongqing, Singapore and Tokyo. The company has a market value of over US$1 billion, with capital injections from powerful venture capital funds including IDG Capital, Sequoia China, ZhenFund, Lenovo Capital and Incubator Group, Green Pine Capital Partners, and CoStone Capital. “There has been an upsurge in AI in recent years and a great demand for the technology in the industrial development of the Greater Bay Area (GBA),” said Dr. Shen. “Gaining a footing in the GBA allows us to collaborate with our clients and create shared value, which will attract more talented individuals to participate in our AI journey. We hope we can go further and develop a comprehensive AI application system to fulfill the smart factory vision of Industrial Revolution 4.0.” 

To do so, Dr. Shen believes that resource allocation and talent cultivation are equally important. “We have a lot of support from CUHK as well as the mainland and Hong Kong governments in terms of advocating and facilitating a variety of exchange projects. We also work closely with the laboratory at CUHK to explore cutting-edge technology. Meanwhile, we provide internships for talented Hong Kong people and help them discover career prospects in the industry.”

“Demand for AI applications is great in the manufacturing industry. Through in-depth cooperation between the industrial sector and CUHK, we hope we can realise and promote the concept of the smart factory. This is one of the ways to make the best use of AI.”

— Dr. SHEN Xiaoyong

Co-Founder & CEO, SmartMore

Research areas: Artificial Intelligence Image Analysis, Video Understanding R&D and Application

Major achievements:

• WAIC Yunfan Award (2021)
• Published over 60 research paper at top-tier international conferences and journals, and holds more than 50 patents
• PhD Graduate of CUHK Computer Science and Engineering