Speakers

Abstract: Power electronic converters, as core devices in modern smart grids, face three major technical challenges in topology derivation, parameter design, and control implementation. The discrepancy between traditional manual solutions and the demands for intelligent operation is increasingly evident, primarily manifested in topology optimization, time-consuming parameter tuning, and a lack of adaptability in control strategies under complex operating conditions. We will discuss the state-of-the-art technologies in AI-based power electronic converters, and elaborate on their specific applications in topology derivation, parameter design, and control implementation. Based on this discussion, we will outline prospects for future research and identify potential challenges in power electronic converters.

Areas of Expertise: Graph Machine Learning, Mobile Computing, Blockchain, and Big Data Analytics

Abstract: In this talk, I will discuss intelligent healthcare data fusion models and their applications combined with graph, blockchain, and multimodal learning. I’ll start by summarizing multimodal data fusion in healthcare engineering and the challenges it faces. Following that, I will quickly go over the key techniques used in our investigations, including multimodal data fusion, blockchain, federated learning, community and node centrality in graphs, graph embedding, graph neural networks, etc. I’ll then focus on applying these techniques to multimodal healthcare data fusion. In particular, I demonstrate our four proposed models, referring to blockchain-based healthcare data fusion, GNN-based healthcare knowledge representation, hierarchical attention-driven multimodal learning, and contrastive learning-based multi-hypergraph learning and diffusion modeling. Then, the architectures, theoretical presentation, realization steps, experiments, and analyses of these models have been carried out. Finally, I summarize our studies and give some perspectives.

Brief Introduction: Hui Liu is Jiangsu Provincially Distinguished Professor at the School of Artificial Intelligence / Future Technology, Nanjing University of Information Science and Technology (NUIST). He is ranked among the top 2% of scientists worldwide by Stanford and placed within the global top 0.05% by ScholarGPS—where he ranks 10th globally in the field of Human Behavior. He serves as Chief Scientist at Guodian Nanjing Automation Co., Ltd. (GDNZ); Chair Professor (by invitation) at the College of Arts, Xi'an University of Architecture and Technology (XAUAT); Guest Professor at the HFU International Institute, Hochschule Osnabrück (Germany); and Researcher at the University of Bremen, Germany. His international academic engagement includes appointments as an Erasmus+ Scholar, recipient of a Mobile Teaching Grant, and selection as a Young European Research Universities Network (YERUN) Scholar with a Mobile Research Award. He was co-responsible for four multi-institutional, nationally and industrially funded German research projects. He pioneered the first intelligent knee bandage capable of real-time human activity recognition, recognized with the Best Paper Award (student author) at BIODEVICES 2019. A subsequent co-authored study on EMG-based facial action unit recognition received the Best Paper Award at BIODEVICES 2025. He has authored or edited over 125 peer-reviewed publications. Eleven of his journal articles are designated ESI Highly Cited Papers (top 1%), including eight Hot Papers (top 0.1%). He has delivered plenary or keynote talks, or served as chairs at more than 20 major international conferences and currently serves as Editor-in-Chief of the Journal of Engineering Research and Sciences, alongside editorial board memberships in multiple SCI journals. Prof Liu's scholarly service and industrial impact have been acknowledged through the Sensors 2023 Outstanding Reviewer Award, the CAMPUSiDEEN Public Choice Award for Smart Sensing and Recognition Technology (2022), and the Fifth Prize in the Hangzhou Innovation & Entrepreneurship Competition for Overseas Talents (2025).

Abstract: In the field of EE, efficient and interpretable time-series analysis holds utmost significance. This keynote presentation centers around the open-access (OA) and highly effective time-series toolkits developed and launched by the speaker in collaboration with their cross-country teams and research partners. The toolkits encompass a wide range of aspects, such as TSFEL (Time-Series Feature Extraction Library, ESI Highly-Cited Paper), TSSEARCH, and Self-Similarity Matrix. These open-source toolkits are designed with a user-friendly approach, following the What You See Is What You Get (WYSIWYG) principle. They offer pre-built functions and algorithms for time-series analysis, which can significantly save the time of researchers and engineers in EE. For instance, they are applicable in voltage and current processing, as well as environmental sensing data analysis. The related contribution also delves into outlier detection. The MS2OD and MMOD algorithms, recognized as ESI Hot Papers, present a novel method for identifying abnormal data points. This is crucial for ensuring the reliability of EE systems. Overall, these OA and powerful toolkits have great potential in energy and electrical engineering applications, facilitating more efficient and accurate time-series analysis and decision-making.