Professor Philip Yu delivers a talk at the RGC public lecture: From Saving Lives to Empowering Learning: Wow, AI is So Close!
Professor Philip Yu Leung-ho is invited by the Research Grants Council (RGC) to give a talk about the rise of AI and its wide range of applications in today’s world.
Professor Yu recaps the development of AI, from artificial intelligence, machine learning, deep learning to generative AI. (Photo credit: The Hong Kong Science Museum)
A group of secondary school students take a photo with the two speakers after the lecture: Professor Philip Yu Leung-ho from EdUHK, far right, back row, and Professor Wong Tsz-wai from HKUST, far right, front row.
Professor Philip Yu Leung-ho, Associate Director of the University Research Facility of Data Science and Artificial Intelligence and former head of the Department of Mathematics and Information Technology (MIT) delivered a talk about the rapid rise of AI on 20 January 2024. Entitled “From Saving Lives to Empowering Learning: Wow, AI is So Close!”, the talk was given as the second instalment in a series of public lectures organised by the Research Grants Council (RGC). Professor Wong Tsz-wai from the Department of Chemical and Biological Engineering of The Hong Kong University of Science and Technology (HKUST) also gave a talk at the same lecture.
Professor Yu opened his talk by giving the audience a brief picture about the development of AI from artificial intelligence, machine learning, deep learning to generative AI. He said the rapid rise of AI in commercial uses, academic research projects and everyday life has been driven by a combination of factors. These included the wide accessibility of open-source tools, enhanced algorithms, availability of big data with labels, and the revolutionary improvements in computing power of standalone or networked devices.
Professor Yu explains that by analysing information available on social media and mapping out the influencing networks of individuals, AI can learn customer preferences and make reasonably good prediction on services and products that match the needs of individual customers.
Generative AI tools are so powerful that they can produce images that are convincingly “real”, challenging the boundary between what is “real” and “fake”. To test how “real” photos produced by Generative AI could look like, Professor Yu asks the audience to judge whether a piece of news is real or fake.
For many years, Professor Yu has worked in research work related to AI and big data analytics, non-parametric inference, ranking methods, time series analysis, financial data analysis, risk management, and statistical trading. In the RGC’s public lecture, Professor Yu used several research projects that he has worked closely with to illustrate AI’s wide range of real-life applications. The professor said AI technology can provide a reliable and rapid computerized medical image diagnosis such as detection of cancer and stroke. It can help medical professionals to make more timely clinical judgments. He said on the education front, educators will find the AI-assisted language learning and assessment model helpful in aiding them to assess quality of picture-cued writing of students.
“Through AI-enabled vector analysis of key hint words like “want to die” (想死), “very tired” (好攰), “huge pressure” (壓力好大) and hopeless (絕望), we can identify comments on social media which are indicative of suicidal risk. This helps concerned professionals to take preventive measures against potential suicidal cases. The high prediction power of AI will ensure it a central role in the commercial world too. For instance, preference learning and prediction function of AI can map out the influencing networks of individuals, hence helping marketers to make the most relevant personalised recommendations to their customers,” Professor Yu said.
Professor Yu believes despite the rapid development of AI, humans will still hold the ethical and legal responsibilities for all decisions made and actions taken in the near future. (Photo credit: The Hong Kong Science Museum)
Around 200 audience members, a noticeable majority of them secondary school students, attended the lecture. When asked by the floor his view on whether AI will replace humankind in the future, Professor Yu agreed that routine and repetitive work types will likely be replaced by AI-assisted technology. He stressed that in the new AI age, authenticity of raw data is rudimentary as predictions made by AI, no matter how powerful it is, hinge on what kind of information it has received. He believes in the nearest future, AI only plays an assistant role when human beings still need to take ethical and legal responsibilities for all the decisions made or actions taken with the assistance from AI.
The public lecture on 20 January 2024 was composed of two talks. The first one was given by Professor Wong Tsz-wai from HKUST. It was about how his research team has developed an AI-based microscope that revolutionises the ways cancer cells are detected before, during, and after surgeries. Compared with the usual one-week processing time, it takes only three minutes for the AI-based microscope to provide high-resolution images of biological tissue samples from a patient that closely resemble images acquired using the highest clinical standard. The new technology enables fast and accurate medical diagnoses for all tissue types.
RGC public lectures aim to arouse public interest in local research developments. Since 2009, numerous leading scholars, including EdUHK academics, have been invited to present their research and achievements to the public. In line with the Government's initiative to promote STEAM (Science, Technology, Engineering, the Arts and Mathematics) education for all, for fun, and for diversity, the RGC is organising a series of public lectures titled “Happy, Healthy, Longevity - AI Can Help” which covers an array of areas including technology, innovation, and health. Held at the Hong Kong Science Museum on 20 January 2024, the second lecture of the series contained the two talks by Professor Yu and Professor Wong.
