Long Cheng

IEEE Fellow, IET Fellow, AAIA Fellow

University of Chinese Academy of Sciences, China

Biography: 

Long Cheng, Researcher at the Institute of Automation, Chinese Academy of Sciences, Professor at the University of Chinese Academy of Sciences, and PhD supervisor. IEEE/IET Fellow. His main research interests include dexterous robot manipulation and imitation learning, embodied intelligence, tactile and force perception and multimodal large models, physical human-robot interaction, and intelligent wearable technologies. Currently appointed as Chair of the IEEE CIS Beijing Chapter, he also serves as Deputy Director of the Technical Committee on Robot, Chinese Association of Automation, Joint Member of the Technical Committee on Control Theory, Chinese Association of Automation, Joint Member of the Cognitive Intelligent Automation, Chinese Association of Automation, Deputy Director of the Technical Committee on Network Science and Engineering and Deputy Director of the Technical Committee on Intelligent Command and Control Systems Engineering, Deputy Director of the Intelligent Vehicles and Robotics Subcommittee of the Analytical Instrument Branch of China Instrument and Control Society, and Vice Chairman of the Beijing Association for Artificial Intelligence. He is an editorial board member of several domestic and international journals, including IEEE/ASME Transactions on Mechatronics, IEEE Transactions on Cybernetics, Science China Information Sciences, and Science China Technological Sciences. He has been awarded the National Science Fund for Distinguished Young Scholars and the Beijing Distinguished Youth Fund. He received the Second Prize of the National Natural Science Award in 2017.

Title: Robot Skill Learning by the Dynamical System Approach

Abstract: A central goal in robotics is to build machines that can fluidly learn complex skills from human demonstration and interact safely and reliably in unstructured environments. While imitation learning has shown significant promise, conventional methods often struggle with a fundamental trade-off between accurately reproducing demonstrated motions and guaranteeing the stability and generalization required for real-world deployment. This talk will present a principled approach to robotic skill learning rooted in the theory of dynamical systems (DS), which models movements not as fixed trajectories, but as vector fields that guide the robot towards a goal with inherent robustness to perturbations. We will trace the evolution of DS-based imitation, from early concepts of movement primitives and Dynamic Movement Primitives (DMPs) to modern techniques that formally address the stability-accuracy dilemma. A key focus will be on the use of diffeomorphic transformations, powered by invertible neural networks, to learn complex, nonlinear skills while providing mathematical guarantees of global stability. This framework enables novel applications in high-precision tasks, such as robotic drawing and forgery detection in signatures, and extends to periodic motions for collaborative tasks like physical rehabilitation through the integration of Neural Liénard systems. Finally, the talk will explore the burgeoning synergy between classical dynamical systems and the new era of foundation models. We will discuss how DS principles can enhance modern AI, offering computationally efficient and physically grounded alternatives to Transformers, such as State Space Models, and improving the physical realism of world models through methods like Variational Information Bottleneck. Conversely, we will look to the future, investigating how large-scale models can help solve long-standing challenges in control theory, such as the automated discovery of global Lyapunov functions. This synthesis charts a path toward a new generation of robotic intelligence that combines the rigor of control theory with the expressive power of deep learning.

Jian Sun

National Leading Talent

Beijing Institute of Technology, China

Biography: 

Jian Sun, Professor and Dean of the School of Automation at Beijing Institute of Technology, and Executive Deputy Director of the State Key Lab of Autonumous Intelligent Unmanned Systems. His main research areas include autonomous intelligent unmanned systems, networked systems, and control system security. He has published over 180 academic papers in journals such as IEEE Transactions and Automatica, and authored 2 academic monographs. He has received one Second Prize of the National Natural Science Award, two First Prizes of the Ministry of Education Natural Science Award, and two Second Prizes of the National Defense Science and Technology Progress Award. In 2019, he was awarded the National Science Fund for Distinguished Young Scholars. He currently serves as a member of the 8th Science and Technology Committee of the Ministry of Education, Deputy Director of the Technical Committee on Control Theory, Chinese Association of Automation, Deputy Director of the Industrial Control System Information Security, Chinese Association of Automation, Deputy Director of the Technical Committee on Swarm Intelligence and Cooperative Control, and Deputy Director of the Technical Committee on Network Science and Engineering. He is also an editorial board member of journals including IEEE/ASME Transactions on Mechatronics, IEEE Transactions on Systems, Man, and Cybernetics: Systems, Science China Information Sciences, Journal of Systems Science and Complexity, Acta Automatica Sinica, and Acta Electronica Sinica.

Title: Data-driven control of networked systems

Abstract: With the development of information technologies, control systems are becoming more intelligent and interconnected. Accurate modeling of a control system has become increasingly difficult. For systems that are difficult to accurately model, traditional model-based control theories and methods are difficult to achieve ideal control performance. Data-driven control refers to the control method of designing controllers based solely on the offline/online data when the mathematical model and parameters of the control system are unknown. Data-driven control methods are independence of precise models and have broad applications. This talk will introduce the recent progress of data-driven control methods for networked systems, including data-driven event-triggered control and self-triggered control, data-driven resilient control under DoS attacks, data-driven self-triggered control based on trajectory prediction, data-driven robust LQG control, and data-driven output regulation of networked systems.

Jian Cao

IEEE Senior Member 

Shanghai Jiao Tong University, China

Biography: 

Jian Cao is a tenured professor and PhD supervisor at Shanghai Jiao Tong University. His main research interests include business process management, intelligent data analysis, service computing and distributed artificial intelligence. He has presided over a number of research projects, including key programs of the National Key R&D Program of China, projects of the National High-tech R&D Program (863 Program), projects of the National Natural Science Foundation of China, and key projects of the Shanghai Municipal Science and Technology Commission. He has received nine provincial and ministerial-level science and technology progress awards, including Special Class, First Class and Second Class honors, and was selected into the Program for New Century Excellent Talents of the Ministry of Education. His research achievements have been applied in industries such as finance, healthcare and tourism. He has published more than 300 academic papers at home and abroad. Currently, he is a Senior Member of IEEE, a Distinguished Member of China Computer Federation (CCF), Director of the Special Committee on Collaborative Computing and Information Services of Shanghai Computer Society, and Deputy Director of the Special Committee on Trusted Intelligent Systems of Shanghai Artificial Intelligence Society.

Title: Peronalized Federated Learning

Abstract: This report explores Personalized Federated Learning (PFL), addressing traditional federated learning's flaws in non-IID data and heterogeneity. It expounds FL’s basics, categorizes PFL methods (global model personalization, direct personalized model learning), introduces the PFLlib simulation platform, and its financial field applications. PFL fits real needs well yet faces efficiency-security balance challenges, requiring industrial collaboration for scaling.

Chengxin Zhang

Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China

Biography: 

Dr. Chengxin Zhang is a Principal Investigator at the Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences. Previously, he served as a Research Assistant Professor at the University of Michigan and as a Postdoctoral Fellow at Yale University. He has authored more than 60 peer-reviewed SCI-indexed publications, earning an h-index of 34 and accumulating over 6,600 citations. His work as first or corresponding author has appeared in prestigious journals such as Nature Methods and Nature Protocols. Dr. Zhang has developed several open-source software tools that have achieved top rankings in community-wide benchmarks, including the Critical Assessment of Structure Prediction (CASP) and the Critical Assessment of Function Annotation (CAFA) challenges.

Title: AI-based Protein and RNA Function Prediction

Abstract: While the biological function of a protein is contingent upon its structure, most function prediction algorithms are unable to directly utilize 3D structures or effectively integrate multi-modal data such as sequence, structure, and interactions, making structure-based functional prediction still challenging. To address this, the speaker developed the COFACTOR algorithm to predict functions from experimentally or computationally derived structures. By performing structural alignment, sequence search, and protein interaction retrieval, the algorithm matches template proteins with functional annotations from databases to infer functions. This approach not only filled a gap in structure-based functional prediction but also significantly improved prediction accuracy. In the "Limited Knowledge Biological Process" category of the third International Critical Assessment of Functional Annotation (CAFA3) from 2016 to 2018, COFACTOR achieved first place. Building upon COFACTOR, the speaker developed a deep learning-based functional prediction algorithm, StarFunc. In CAFA5 in 2024, which involved 1,625 teams from 96 countries, StarFunc ranked fifth. Recently, the algorithm has been extended to the field of RNA, resulting in the first Gene Ontology (GO) function prediction model using RNA sequence as the only input. The accuracy substantially surpasses that of existing state-of-the-art methods. These algorithms developed by the speaker have been widely applied to the structural and functional annotation of whole genomes in humans and pathogenic microorganisms. Notably, the deep learning-based functional annotations of human proteins have been incorporated into the neXtProt database, the official repository of the Human Proteome Organization (HUPO).

Qi Zhang

Tongji University‌, China

Biography: 

Qi Zhang, a tenure-track associate professor at the School of Computer Science and Technology, Tongji University, holds a Ph.D. in Computer Science and Technology from Beijing Institute of Technology and a Ph.D. in Analytics from the University of Technology Sydney. He has led and participated in more than 10 national, provincial, and ministerial projects. His main research interests lie in multimodal information processing and AI for Science, specifically including brain visual coding and decoding, time series analysis, and multimodal learning. He has published over 80 papers in top international journals and conferences in the fields of artificial intelligence and data science, holds 5 invention patents, and serves as the deputy secretary-general of the CSIG Youth Working Committee and the Computer Vision Specialty Committee of the Shanghai Computer Society, and the YOCSEF Shanghai Academic AC.

Title: Multimodal brain-vision encoding and decoding

Abstract: Multimodal brain-vision encoding and decoding aims to utilize various brain functional data (such as functional magnetic resonance imaging) to build a bidirectional bridge between artificial intelligence and human brain vision mechanisms. On the one hand, this research simulates the visual information processing process of the human brain under different modalities through the "encoding" model; on the other hand, it reconstructs or classifies the visual stimuli seen by the human eye directly from the collected brain signals through the "decoding" model. This interdisciplinary field not only provides a new paradigm for exploring the principles of brain visual cognition but also holds broad application prospects in the fields of brain-computer interfaces and brain-inspired intelligence.

Zhen Yang

Jiangxi Science and Technology Normal University, China

Biography: 

Zhen Yang, Professor at Jiangxi Science and Technology Normal University, Vice Dean of the Graduate School, and Deputy Director of the Jiangxi Provincial Key Laboratory of Advanced Electronic Materials and Devices. He is a master's supervisor and a recipient of the Jiangxi Provincial Outstanding Youth Fund. He holds a Ph.D. from Shanghai Jiao Tong University. His main academic research areas include artificial intelligence, pattern recognition and image processing, computer vision, and machine learning. His specific research directions are hyperspectral image analysis and processing, crop image classification, agricultural pest target detection and recognition, and human action analysis. He has published over 100 papers in mainstream academic journals and conferences at home and abroad. He has also led two regional projects of the National Natural Science Foundation of China and two projects of the Jiangxi Provincial Natural Science Foundation. He has won the second prize of the Jiangxi Provincial Teaching Achievement Award.

Title: Random Shuffling Data for Hyperspectral Image Classificatio

Abstract: Hyperspectral images (HSIs) are characterized by hundreds of spectral bands. The goal of HSI is to associate the pixel with a corresponding category label by analyzing subtle differences in the spectrum. Due to their excellent local context modeling capabilities, Convolutional Neural Network (CNN)-based and Transformer-based methods are often adopted to complete the classification task. To verify whether the patch-data-based CNN methods depend on the homogeneity of patch data during the training process in HSI classification, we designed a random shuffling and the Patch-Level Spatial Perturbation (PLSP) data strategy to disrupt the data homogeneity of the patch data, which is randomly assigning the pixels from the original dataset to other positions to form a new dataset. Based on this random shuffling strategy, we propose a sub-branch to extract features on the reconstructed dataset and fuse the loss rates (RFL). In addition, a novel wavelet-enhanced spatial model framework for hyperspectral image classification is proposed, and it achieves high classification accuracy on public datasets such as Indian Pines (IP), Pavia University (PU), Salinas (SA), and Longkou (LK).

Tao Zhang

Shanghai Jiao Tong University, China

Biography: 

Tao Zhang is an Associate Professor at the School of Automation and Intelligent Sensing, Shanghai Jiao Tong University, who is also an Associate Editor of IEEE Transactions on Image Processing and the Secretary-General of the Shanghai Chapter of IEEE GRSS. His research focuses on pattern recognition, SAR image scene interpretation, and land cover/landform change detection. He has led projects including the National Natural Science Foundation of China (NSFC) Young Scientist Fund, China Postdoctoral Science Foundation, and National Key Laboratory Open Fund, and participated in multiple research initiatives including NSFC Major/Key/General Projects. He has published over 70 SCI-indexed papers in renowned domestic and international journals (including 4 ESI highly cited papers) and holds 5 national invention patents (applied/authorized), which contribute to many awards, including the First Prize for Scientific and Technological Progress from the Shanghai Automation Society, the First Prize for Service Industry Technological Innovation from the China General Chamber of Commerce, the Second Prize for Teaching Achievement from the Shanghai Computer Society and so on.

Title: SAR ship detection with the polarimetric information

Abstract: Synthetic aperture radar (SAR) is a microwave imaging system with capabilities to image the earth surface day-and-night under clear and cloudy conditions. In recent years, ship detection in SAR data has attracted a great deal of attention to help governments deal with marine-traffic control, illegal fishing, and immigration enforcement, among others. Compared to single-polarimetric mode data, multipolarimetric mode, especially the quad-polarimetric mode data (i.e., PolSAR), contains both backscattering intensity and phase information and allows us to explore the scattering differences between target and clutter. To enhance the detection precise of SAR ship targets, in our current works, different backscattering matrices related to the use of polarimetric information are developed, such as [P] and [CP]. Experiments carried out on many different datasets have also demonstrated their effectiveness. In particular, the target-clutter ratio of weak ship targets can be significantly improved.

Mou Chen

IEEE Fellow, IET Fellow, CAA Fellow

Nanjing University of Aeronautics and Astronautics, China

Biography: 

Mou Chen, an IEEE Fellow, IET Fellow and a CAA Fellow, serves as the Dean of the College of Automation Engineering at Nanjing University of Aeronautics and Astronautics. He was the recipient of the National Science Fund for Distinguished Young Scholars in 2018, was selected into the National “Hundred-Thousand-Ten Thousand Talents Project” in 2019, and was included in the “New Century Excellent Talents Support Program” of the Ministry of Education in 2011. Currently, he serves as an editorial board member of several SCI-indexed English journals, such asIEEE Trans. Cybernetic,IEEE/ASME Trans. Mechatronics,IEEE Trans.CS II: Express Briefs, etc, and also serves as an editorial board member of Chinese journals including Science China: Information Sciences, Acta Aeronautica et Astronautica Sinica, Acta Automatica Sinica, Control Theory & Applications, etc. He has successively wonthe Second Prize of the National Natural Science Award (ranked second),the First Prize of the Jiangsu Provincial Science and Technology Award (ranked first), the First Jiangsu Provincial Outstanding Contribution Award for Young Scientists and Technologists,the 2First Prize of Provincial and Ministerial Award (rankedfirst), and 2 Second Prizes of the National Defense Science and Technology Progress Award (ranked first). He has applied for and been authorized more than50 invention patents. He has published 3 monographs in Chinese and English and has published more than 200 academic papers.

Title: Safe and Intelligent Control Technology for High-Maneuverability Unmanned Aerial Vehicles

Abstract: High-maneuverability unmanned aerial vehicles (UAVs) exhibit characteristics such as unsteady aerodynamics, strong nonlinearity, severe dynamic uncertainties, strong coupling, and significant disturbances. These features may invalidate traditional linearized flight control methods, thereby seriously threatening the flight safety of UAVs. In response to the flight control requirements of large flight envelopes and multi-mission operations for high-maneuverability UAVs, this report mainly introduces research achievements in four aspects: the calculation and prediction of safety boundaries for high-maneuverability UAVs, multi-approximator cooperative control, robust constrained control, and multi-index constrained control. In addition, the application of the developed intelligent flight control methods in UAVs is discussed, and an outlook on future research directions is provided.

Huaxiang Zhang

IEEE Member

Aerospace Information Technology University, China

Biography: 

Huaxiang Zhang, Doctor of Engineering from Shanghai Jiao Tong University, Level-2 Professor, Doctoral Supervisor, Taishan Scholar Distinguished Professor, Young and Middle-Aged Expert with Outstanding Contributions in Shandong Province, High-end Talent of Shandong Think Tank, Top-notch Professional and Technical Talent in Jinan City, Dongyue Scholar of Shandong Normal University, etc. He has successively served as Dean of the School of Information Science and Engineering, Director of the Science and Technology Department, and Director (Dean) of the Science and Technology Department (Institute of Science and Technology) of Shandong Normal University, Member of the Party Committee and Vice President of Shandong Jiaotong University, and is currently a Member of the Interim Party Committee and Deputy Director of the Connotation Construction Working Task Force of Aerospace Information Technology University. He is a Senior Member of the China Computer Federation (CCF), IEEE Member, ACM Member, Member of the Chinese Special Committee on Machine Learning, Member of the Chinese Special Committee on Artificial Intelligence and Pattern Recognition, Member of the Special Committee on Big Data and Artificial Intelligence of the Chinese Society for Industrial and Applied Mathematics, Vice Chairman of the Shandong Artificial Intelligence Society, etc. He also serves as a review expert for national, provincial and municipal natural science funds as well as teaching and science and technology awards, and a specially invited reviewer for many well-known domestic and international academic journals.

Title: TBD

Abstract: TBD