Hong Kong is a place having the world’s top-ranked healthcare systems and infrastructures, yet also a place faced with challenging demand of healthcare services due to aging population and rising cancer incidence. In 1993, Hong Kong has 9% population at aged 65 or above, and this percentage is expected to reach 26% in 2031. Innovating medical technologies for advanced solutions to overcome doctor shortage and reduce medical costs have been top priorities in recent government initiatives and policies. One current major strategy of hospital authority, which manages all public hospitals in Hong Kong, is to promote Smart Hospitals with intelligent services for sustainable development.

Artificial intelligence (AI) in healthcare has witnessed exploding research enthusiasm globally and demonstrated promising success stories in recent years. A wide scope of medical applications are suitable to be better tackled with the help of intelligent machines that can speed up massive data analysis, reduce manual repetitive workload and facilitate informed decision makings. Currently, the investigated topics include AI for medical image diagnosis, electronic health record interpretation, genomics data analysis, protein structure prediction, drug development, medical robotics, remote health monitoring, smart wearable/portable devices, ambient assisted living, mental health and rehabilitation, medical education and so on. For examples, at present global fight against COVID-19, AI systems are developed to automatically assist detection of infections in lung Computed Tomography (CT) images, which can facilitate timely patient triage and management during pandemic. In DeepMind’s latest news, their developed AI model called AlphaFold2 can accurately predict 3D structure of human proteins, being a breakthrough shedding lights on a 50-year-old grand challenge in biology. Basically, it will be no surprise to see AI transforming the future of healthcare in many known and unknown ways by interaction with clinical practice, biomedical knowledge and life science. 

Alongside, there exist important remaining issues and open challenges to be considered, discussed and addressed. First of all, existing healthcare AI systems have limited collaboration with clinicians. Human-factors with valuable domain expertise are outside the loop of AI technical development and deployment. How to strengthen human-centered AI is a crucial next-step to fulfil the ultimate value in clinical context. Second, limited interpretability of AI networks is a fundamental machine learning challenge, which prohibits transparency of explainable reasons that lead to critical decision makings in healthcare domains. Third, translations of AI to clinical trials have to overcome many real-world data-oriented obstacles, including existence of low-quality data due to complex acquisition situation, data distribution shift across different hospitals using various machines, unstructured data without following standardized clinical workflow, etc. Fourth, ethical use of AI in health is under exploration regarding concerns on data privacy, model biases and safety. The World Health Organization (WHO) has also recently released a guidance on ethics and governance of artificial intelligence for health.

The challenges are in fact tightly coupled and the solutions have to rely on interdisciplinary efforts between engineering and medical experts. CUHK has unique strong Faculty of Engineering and Faculty of Medicine, and the created multidisciplinary collaborations have been long-lasting and fruitful. Up to now, a diversity of research and development of AI techniques for healthcare applications have been led by engineering faculties across different specialized disciplines. The on-going projects cover almost all above-mentioned topics with multi-aspect achievements on high-impact publications, prestigious research awards and industrial technical transfer. Enormous opportunities have also been provided to Engineering students for training, practice and innovations of cutting-edge technologies and multi-disciplinary knowledge, to prepare for meeting the foreseeable increasing demand for AI healthcare talents in Hong Kong and the globe.

Prof. Dou Qi

Department of Computer Science and Engineering