IEEE SMC Distinguished Lecture given by Prof Mengchu Zhou

Join Prof. MengChu Zhou for an IEEE SMC Distinguished Lecture on student-centred knowledge distillation and its AI applications. The IEEE Systems, Man, and Cybernetics Society (SMC) Portsmouth Chapter is pleased to invite you to an IEEE SMC Distinguished Lecture delivered by Prof. MengChu Zhou, Distinguished Professor of Electrical and Computer Engineering at the New Jersey Institute of Technology (NJIT), USA. The lecture will explore recent advances in knowledge distillation, with particular emphasis on student-centred approaches inspired by human educational principles, control systems, curriculum learning, and their applications to computer vision. The event will be delivered in a hybrid format, and participants may attend either in person at the University of Portsmouth or online. Lecture Title A Multi-discipline Approach to Advancing Knowledge Distillation and its Applications Abstract Knowledge distillation is among the core technologies in Artificial Intelligence, and its deployment enables users to benefit from many AI applications on resource-constrained end devices such as mobile phones. Existing studies on knowledge distillation mainly focus on teacher-centred methods, in which the teacher network is trained according to its own standards before transferring the learned knowledge to a student network. However, due to differences in network structures between teacher and student networks, the knowledge learned by the teacher may not necessarily be what the student requires. Inspired by human educational wisdom, we propose a Student-Centered Distillation (SCD) method that enables the teacher network to adjust its knowledge transfer according to the student's actual needs. The approach incorporates principles inspired by human education. For example, the teacher network identifies and learns the knowledge required by the student network using the validation set and subsequently transfers this knowledge to the student through the training set. To address challenges including knowledge deficiency, hard-sample learning, and knowledge forgetting during the student's learning process, we introduce and improve Proportional-Integral-Derivative (PID) control algorithms to identify the knowledge currently required by the student network. Furthermore, a curriculum-learning-based fuzzy strategy is integrated with the proposed PID control algorithm, enabling the student network to actively focus on challenging samples. Experimental results demonstrate that SCD outperforms existing teacher-centred knowledge distillation methods across multiple computer vision tasks, ranging from object detection to image segmentation. About the Speaker Prof. MengChu Zhou is a Distinguished Professor of Electrical and Computer Engineering at the New Jersey Institute of Technology (NJIT), USA. He received his B.S. degree in Control Engineering from Nanjing University of Science and Technology in 1983, his M.S. degree in Automatic Control from Beijing Institute of Technology in 1986, and his Ph.D. degree in Computer and Systems Engineering from Rensselaer Polytechnic Institute in 1990. His research interests include Petri nets, intelligent automation, Internet of Things, big data, web services, and intelligent transportation. He has published more than 800 works, including 12 books and over 500 journal papers, and holds 12 patents. Prof. Zhou is the founding Editor of the IEEE Press Book Series on Systems Science and Engineering and Editor-in-Chief of the IEEE/CAA Journal of Automatica Sinica. His honours include the Humboldt Research Award, the IEEE SMC Society Franklin V. Taylor Memorial Award, and the IEEE SMC Society Norbert Wiener Award. He currently serves as Vice President for Conferences and Meetings of the IEEE Systems, Man, and Cybernetics Society and is a Fellow of IEEE, IFAC, AAAS, and CAA. Who Should Attend? The lecture will be of particular interest to: • Researchers and academics working in Artificial Intelligence and machine learning • PhD and postgraduate students • Computer vision researchers • Researchers interested in knowledge distillation and efficient deep learning • Researchers working on intelligent systems, control systems, and edge AI • IEEE members and the wider research community Attendance is free, but registration is required. Online Attendance This event is available both in person and online. Online participants can join on Teams. In-person Venue Room 0.34, Portland Building University of Portsmouth Portland Street Portsmouth PO1 3AH