CUHK’s Department of Systems Engineering and Engineering Management Professor Cheng Hong and postdoctoral researcher Dr Sun Xiangguo were awarded the Best Paper Award (Research Track) from ACM SIGKDD Conference on Knowledge Discovery and Data Mining 2023.  Launched in 1989, SIGKDD is the premier international conference for data mining researchers and practitioners from academia, industrial communities, and the government to share exchange ideas, research results and experiences. This seminal work on graph prompt learning is awarded for its innovative idea and methodology.  This is the first time that researchers from Hong Kong and the mainland have received this award from SIGKDD.

The paper, “All in One: Multi-task Prompting for Graph Neural Networks”, presents a novel prompt learning method for graph neural networks in the multi-task background, which can fill the gap between pretrained models and various graph learning tasks, and lay the foundation for the development of large graph models.  The proposed prompting framework for large-scale pre-trained graph models can effectively leverage prior graph knowledge to relieve the lack of graph annotations from specific applications, catalysing the creation of versatile graph learning models.

Professor Cheng received her PhD in Computer Science from the University of Illinois at Urbana-Champaign in 2008.  Her research interests lie in data mining and data management with an emphasis on large-scale graph learning and management, data mining algorithms, and their applications.  She received the Certificate of Recognition for the 2009 SIGKDD Doctoral Dissertation Award, and the 2010 Vice-Chancellor’s Exemplary Teaching Award of CUHK.