On July 22, Professor Witold Pedrycz, Fellow of the Royal Society of Canada, IEEE Life Fellow, and Professor at the University of Alberta, attended Guangxi University’s “Overseas Masters Forum”. He presented an academic report titled “New Pursuits of Machine Learning: Data-Knowledge Environment and Knowledge Landmarks” in the conference room of the School of Mathematics and Information Science. The event was chaired by Professor Xiao Jianzhuang, Vice President of Guangxi University.

In his presentation, Professor Pedrycz addressed core challenges in purely data-driven machine learning, including “data blindness”, the black-box nature of models, and data sparsity. He proposed a novel paradigm for constructing a “Data-Knowledge Environment”, emphasizing that incorporating domain knowledge—such as physical laws and symbolic knowledge—is key to overcoming current limitations. He introduced the concept of “knowledge landmarks” as a core methodological tool, explaining how key knowledge anchors can be extracted from data and optimized, and how these landmarks can be integrated into modeling frameworks such as Gaussian Processes to enable knowledge-guided learning. Additionally, Professor Pedrycz discussed practical approaches to leveraging both data and knowledge in rule-based model design. Through case studies, he demonstrated the advantages of data-knowledge collaborative models, and summarized the key paths, challenges, and opportunities in moving toward “neuro-symbolic machine learning”.

During the session, Professor Pedrycz engaged with attendees and provided detailed responses to their questions.

The event was attended by over 40 faculty and student representatives from the School of Electrical Engineering, School of Civil Engineering and Architecture, School of Computer and Electronic Information, and School of Mathematics and Information Science.