![]() | Prof. Xiang ChengPeking University, China Bio: He is a Boya Distinguished Professor at Peking University, an IEEE Fellow, and a National Outstanding Young Scientist. A Fellow of the Chinese Institute of Electronics, the Chinese Association of Automation, and the Chinese Institute of Communications, he has published over 160 papers in Nature-branded and IEEE journals, holds 40 granted patents, and has received the Second Prize of the National Natural Science Award along with 11 provincial- and ministerial-level honors. He currently serves as Editor-in-Chief of Advances in Signal Processing (Springer Nature) and Chair of the Technical Committee on Networked Intelligence of the Chinese Association of Automation. Speech Title: WiFo: A Family of Wireless Communication-Sensing Foundation Models, from Unimodal to Multimodal Abstract: This report introduces Synesthesia of Machine (SoM), a new concept and architecture proposed by our team, and reviews the latest progress of the WiFo series of wireless communication-sensing foundation models — from RF-modality pre-training to multimodal design for integrated sensing and communication. By large-scale pre-training on communication and multimodal sensing, WiFo achieves native modeling and unified representation of sensing and communication information, offering general and transferable capabilities for SoM transceiver design. We further present the first hardware system with an embedded wireless foundation model, followed by a summary and future outlook. Papers about above topics have been published in TMI,ICCV,ECCV,MICCAI in recent years. |
![]() | Prof. Zhaoshui HeGuangdong University of Technology (GDUT), China Bio: Zhaoshui He is a full professor with Guangdong University of Technology, served as the chief Scienctist of National Key Research and Development Program, and selected in Ten Thousand Talents Program for National Science and Technology Innovation of China. He also heads the director of Research Center of Guangdong Provincial engineering technologies and deputy director of Key Laboratory of National Ministry of Education. His research interests focus on intelligent information processing and smart healthcare. He had published over 100 papers in high level journals such as IEEE Trans PAMI. He applied over 60 invention patents; Received multiple provincial and ministerial level scientific awards, including the second prize of the National Natural Science Awards; supported by a series of foundations, including the National Key R&D Program of China and the National Natural Science Foundation of China for Excellent Young Scholars. Speech Title: Key Technologies and System of Acupuncture robot Abstract: Acupuncture is one of Chinese treasures. It is very efficient for various diseases but with no side effects so that it is recommended by the World Health Organization and has become popular around the world. For this reasion, there is a growing demand for acupuncture but there are an insufficient number of senior acupuncturists. To solve this problem, we have carried out the development of acupuncture robot. This talk mainly illustrates four key technologies and four advances of acupuncture robot as follows: 1. digitization of acupuncture manipulation and experience automation based on database and human acupoint knowledge graph; 2. acupoint automatic recognition and precise position based on multimodal image deep learning; 3. state perception and control for the accurate and safe manipulation of acupuncture robots; 4. intelligent evaluation and safety verification of acupuncture curative effect based on nonlinear modeling by deep neural network. |
![]() | Prof. Keke HuangCentral South University, China Bio: Keke Huang is a Professor at Central South University, where he serves as Associate Dean of the School of Automation and concurrently as Associate Dean of the National Graduate College for Elite Engineers. He is a Chang Jiang Scholar Distinguished Professor and has also been recognized as a Leading Talent in Scientific and Technological Innovation and a Distinguished Young Scholar of Hunan Province. His research focuses on industrial intelligence and systems. He has published over 100 papers in leading journals and conferences, including IEEE Transactions, IFAC flagship journals, and CCF-A conferences, and holds more than 30 invention patents granted in China and the United States. He has led over 10 national research projects, including projects under the National Key R&D Program of China, research tasks under national science and technology major projects, and projects funded by the National Natural Science Foundation of China. His awards include the Second Prize of the National Science and Technology Progress Award, the First Prize of the Hunan Provincial Science and Technology Progress Award, the Outstanding Young Scientist and Technologist Award of the Nonferrous Metals Society of China, and the IEEE TCSDM Young Professional Award. He serves on the editorial boards of IEEE TASE, IEEE SMCM, and IET CPS, and as Vice Chair of the Academic Committee on Automation of the Nonferrous Metals Society of China. Speech Title: Advanced Process Control and Optimization Industrial Software for Non-Ferrous Metallurgy with Multiphase and Multiphysics Coupling Abstract: Nonferrous metallurgical processes feature intricate multiphase and multiphysics coupling, where the reaction atmosphere exerts a decisive influence on product morphology, conversion efficiency, and resource consumption. Industrial software serves as a core enabler for the regulation of reaction atmospheres. Nevertheless, conventional advanced process control software falls short in capturing coupled multi-reaction interactions and multiphase–multiphysics effects when handling complex raw materials. Targeting the requirements for high-efficiency, low-carbon and intelligent autonomous operation of nonferrous metallurgical workflows, this presentation identifies major bottlenecks in control methodologies, software engineering, as well as deployment and field implementation. Further, it elaborates on core technologies including precise characterization and control of multiphase reaction systems, intelligent model libraries and agile development platforms, and intelligent software deployment within dynamic heterogeneous environments. Finally, real-world industrial case studies verify that the proposed industrial software achieves improved control precision, cost reduction, efficiency enhancement, and autonomous intelligent operation of metallurgical processes. |
![]() | Prof. Xiaochun CaoSun Yat-sen University, China Bio: Xiaochun Cao, Dean of the School of Cyber Science and Engineering, Sun Yat-sen University; Member of the University Academic Committee. His applied basic research focuses on the intersection of artificial intelligence and cyberspace security. He is a recipient of the Outstanding Youth Science Fund (continued support under the Excellent Young Scientists Fund A programme), the National Science Fund for Distinguished Young Scholars, and was selected for the Young Top-Notch Talent Support Programme (Ten Thousand Talents Plan). He has won one provincial/ministerial Special Prize, two First Prizes and two Second Prizes. Six of his supervised graduate dissertations were recognised as outstanding doctoral or master’s dissertations by the Chinese Institute of Electronics, CCF and the Chinese Academy of Sciences; eight of his former students have been selected for national-level talent programmes. Speech Title: AI Security / Artificial Intelligence Security Abstract: Most mainstream artificial intelligence models nowadays rely on data-driven function fitting, and their robustness faces twofold challenges: complex disturbances in natural environments such as rain, mist, haze and fog, as well as deliberately designed adversarial attacks against the digital-physical world. This talk will present our team’s recent advances in AI robustness and vulnerability, multimodal fusion perception and security, cognitive-domain situation awareness and security, and full-lifecycle security assessment. The content covers traditional vulnerability issues including adversarial examples, backdoor attacks and privacy risks in both digital and physical domains, as well as our technical explorations in attribution mechanism modelling, counterfactual generation, and assistive positive perturbation. |