Prof. Huanhuan Chen

University of Science and Technology of China, China

Huanhuan Chen, IEEE Fellow, is a Professor at the School of Computer Science, University of Science and Technology of China. He is the leading talent of the National High-Level Talents Special Support Program (National “Ten Thousand Talents Plan”). He has published over 200 papers in prestigious journals and major conferences in the field of artificial intelligence. He is a recipient of the Wu Wenjun Artificial Intelligence Science and Technology Progress Award, the First Prize of Anhui Province Teaching Achievement Award, and the “Excellent Supervisor Award” of the Chinese Academy of Sciences. His research achievements have been recognized with multiple honors, including the Second Prize of the Ministry of Education Natural Science Award, the International Neural Network Society Young Scientist Award, the IEEE Transactions on Neural Networks Outstanding Paper Award, the IEEE Computational Intelligence Society Outstanding PhD Dissertation Award, and the British Computer Society Distinguished PhD Dissertation Award. As a principal investigator, he has led major national research projects such as the “Science and Technology Innovation 2030 – New Generation Artificial Intelligence” initiative, key projects under the National Key Research and Development Program, major program and key program grants from the National Natural Science Foundation of China, general program grants, and collaborative research projects with the Royal Society of the United Kingdom.

Speech Title: Causal Learning and its Applications

Abstract:In recent years, causal learning has gradually emerged as a major research hotspot in the field of artificial intelligence, attracting increasing attention from both academia and industry. Unlike traditional data-driven machine learning methods that mainly focus on correlations, causal learning aims to uncover the underlying cause-and-effect relationships hidden within data, thereby enabling more reliable reasoning, stronger generalization ability, and more trustworthy decision-making. As artificial intelligence systems are increasingly applied to high-stakes domains such as healthcare, finance, and intelligent manufacturing, the need for robust, interpretable, and transferable models has made causal learning an important research direction for next-generation AI. This talk will introduce key topics related to causal discovery, causal inference, and causal decision-making, covering both fundamental theories and recent technological advances. It will provide a comprehensive overview of the latest developments in causal learning, large-scale causal structure learning, treatment effect estimation, and causality-enhanced large language models. The presentation will also discuss current challenges and future research opportunities in building trustworthy and human-centered artificial intelligence systems with causal learning techniques.

Prof. Xuguang Lan

Xi'an Jiaotong University, China

Xuguang Lan is a professor and Ph.D. supervisor. He is a recipient of the National Science Fund for Distinguished Young Scholars. Currently, he is a professor at the College of Artificial Intelligence, Xi'an Jiaotong University.

His research interests cover computer vision, statistical learning and pattern recognition, deep reinforcement learning and robot learning, multi‑agent games and decision‑making, as well as human‑robot collaboration.

He serves as a Council Member of the Chinese Association of Automation, Director of the Trico-Robot Committee of the Chinese Association of Automation, Council Member and Deputy Secretary-General of the Chinese Society, Vice Director of the "Cognitive Systems and Information Processing" Committee of the Chinese Association for Artificial Intelligence, Vice Director of the "Intelligent Unmanned System Modelling and Simulation" Committee of the Chinese Association for System Simulation, and Vice Director of the Chinese Society for Electrical Engineering, and is a Senior Member of IEEE.

Dr. Xun Xu

A*STAR, Singapore

Dr. Xun Xu is a Senior Scientist II and Deputy Research Director for the Semiconductor Directorate at A*STAR’s Institute for Advanced Intelligence and Computing, Singapore. His research focuses on resource-efficient AI, robust learning, and agentic visual reasoning for semiconductor inspection and 3D vision with publications in IEEE TPAMI, NeurIPS, CVPR, etc. He received his PhD from Queen Mary University of London in 2016. He has served as an Area Chair for NeurIPS, ICML, CVPR, ACM Multimedia, SIGKDD, etc. He received the Outstanding Area Chair Award at ACM Multimedia 2024, best academic paper award in EPTC 2022, A*STAR AI Niche Skill Allowance, etc. He is an IEEE Senior Member.

Speech Title: Efficiency, Robustness and Reasoning for Industrial Visual Inspections

Abstract:Industrial visual inspection must detect rare defects reliably despite limited annotations, changing production conditions, and demands for interpretable decisions. This talk presents advances in efficient learning, model robustness, and vision-language reasoning to address these challenges. We explore how foundation models, semi-supervised learning, and adaptation methods reduce labeling requirements and improve performance under distribution shifts. We then examine multimodal models that combine defect localization with semantic explanations and refined reasoning. Semiconductor case studies illustrate applications in CAD-guided defect detection, SEM image segmentation, and automatic defect classification. The talk concludes with research directions toward more data-efficient, reliable, and explainable industrial inspection.

Professor-level Senior Engineer Yue Zhao

National Key Laboratory of Security Communication, China

Dr. Zhao Yue is a professor-level senior engineer and serves as chief technical officer of the National Key Laboratory of Security Communication. Holding a Ph.D. in Engineering, he is a doctoral supervisor and appointed as a distinguished lecture professor under the Chongqing Bayu Scholars Program. Recognized as a Leading Technological Innovation Talent of Sichuan Province, he also serves as a joint graduate supervisor at CETC, Shanghai Jiao Tong University, Sichuan University, Southwest Jiao Tong University, and Chongqing University of Posts and Telecommunications.

Dr. Zhao Yue has maintained long-term technical leadership in the core fields of confidential communication and information security. Over the past five years, he has led more than 10 major national programs, with total approved research funding exceeding RMB 140 million. As a principal investigator, his work bridges fundamental research and high-impact engineering deployment: he has published 40+ peer-reviewed papers as first or corresponding author in top-tier venues including IEEE Transactions on Knowledge and Data Engineering (TKDE), IEEE Transactions on Mobile Computing (TMC) and Chinese Journal of Electronics (CJE), authored 3 academic monographs, and holds high-value 20+ granted national invention patents and 8 registered software copyrights, all as the first right holder.

Speech Title: When Cybersecurity Knowledge Cannot Agree: Intelligent Semantic Collaboration for End-to-End Multi-source Knowledge Fusion in Cybersecurity

Abstract:The keynote presentation is a systematic academic review based on 168 primary studies (from 3,967 records), defining semantic collaboration as a complete pipeline covering four technical dimensions: Semantic Representation, Semantic Extraction, Semantic Fusion, and Semantic Protection. Its core contribution is proposing a unified taxonomy and process model for multi-source cybersecurity knowledge integration, and thoroughly discussing key technologies including encrypted graph storage, private retrieval, and dynamic updates throughout the cybersecurity data lifecycle. The presentation points out that trustworthy cybersecurity data requires co-optimizing accuracy, scalability, interpretability, adaptability, and privacy. Meanwhile, it focuses frontier research on technical bottlenecks such as data uncertainty, temporal evolution issues, privacy leakage risks, and trust deficits across module boundaries of the semantic collaboration pipeline, emphasizing that the entire technical chain should be evaluated under realistic cross-domain conditions rather than individual components, thereby providing theoretical support and practical pathways for building end-to-end trustworthy semantic collaboration frameworks in cybersecurity.