In Focus - Issue 33 (Spring 2021)

mart bus schedules, taxi dispatching, typhoon warnings, medical diagnoses, and ntech are all set to be part of everyday life in Hong Kong soon, thanks to the development of a next-generation arti cial intelligence (AI) computing cloud platform that will drive and deliver Hong Kong’s smart city applications. One of the key people helping to implement this dynamic integrated “IT brain” is Prof. CHEN Kai, Computer Science and Engineering. He is working together with Hong Kong government departments to set up the technology infrastructure to realize the vision of smart mobility, living, environment, people, government, and economy encompassed in the government’s Smart City Blueprint, announced in . The high-performance hub will also enable researchers and practitioners to participate in data collaboration through inter-city knowledge sharing. Prof. Chen, who is still only and identi es with the “smart”-technology generation – was far from a city boy when he was younger, originally hailing from a village of just seven families in landlocked Anhui province in eastern China. There, he had to walk kilometers to attend school in a nearby town, only venturing to a major city – Shanghai – for the rst time as a teenager. That one visit, though, was enough to show him another world besides mountains, and make him realize that he wanted to be part of it. “I learned that the world was colorful and it inspired me to go out and explore when I grew up,” he said. Now Prof. Chen is using that curiosity about the wider world to take urban life to the next level. In July , he became the youngest academic yet to receive prestigious Research Grants Council (RGC) Theme-based Research Scheme funding when a budget of HK$ million was approved for the smart city infrastructure IT platform. He is serving as project coordinator in the collaborative project with other universities to generate a high- performance distributed machine learning framework. The framework seeks to e ciently handle and make accurate predictions from graph-based streaming data to support smart city applications for key areas such as transportation optimization, urban planning, and crowd sensing. To do this, it will need to overcome the current challenges including data scarcity, algorithm limitations, and computing power ine ciency. One proposed solution to the rst two issues is to devise inter-city transfer learning algorithms, which will enable knowledge learned from other cities with rich data sources to be added to the Hong Kong model and vice versa. Meanwhile, high-performance distributed AI computing architecture to support such deep learning and transfer learning for large graph streaming data, including the adoption of remote direct memory access (RDMA), aims to achieve high-throughput, low-latency communications to improve cluster computing e ciency. S 17 IN FOCUS Prof. Chen Kai and his research team are driving forward the development of a comprehensive smar t urban environment through their advanced high-performance AI computing cloud platform

RkJQdWJsaXNoZXIy NDk5Njg=