Keynotes

Abstract: Three main thrusts behind the success of AI are (1) VLSI (2) big data (with LLM) and deep learning neural networks (NNs). In particular, today’s data-driven AI2.0 depends on NNs to help distill massive Data to Knowledge (D2K). Thus, this talk will start with an overview on the technological evolution from MLP (NN1.0), CNN (NN2.0) and transformers (NN3.0). Thereafter, our attention will be shifted to three far-reaching issues regarding the future AI. First, Neural Architectural Search (NAS) represents a major on-going research front on boosting the power of deep learning. There is abundant biological/empirical evidence suggesting that different neurons possess different types/amounts of information, some relevant but some not. This distinction plays a vital role in pruning of neurons in our XNAS learning paradigm, aiming at structural optimization of MLPs, CNNs, and transformer networks. Second, we note that AI2.0 can do well on certain tasks but poorly on others. Consequently, a new learning paradigm becomes imperative to help inch closer towards Artificial General Intelligence (AGI). A short-term solution could be some sort of seamless integration merging AI together with and the rich domain knowledge. However, and idealistically speaking, a long-term solution would be a revolutionary learning paradigm to facilitate logical reasoning, scientific inference and, even further, inferring new knowledge from old knowledge (i.e. K2K). Finally, we note a new standing revised recently by California’s Board of Education that ``data science as an alternative to algebra diverts students from obtaining mathematical skills required for a broad range of careers”. It is a timely endorsement of this talk’s main theme that ``Math is the Cornerstone of AI”.

Bio: S.Y. Kung, Life Fellow of IEEE, is a Professor of Electrical and Computer Engineering at the Princeton University. His research areas include VLSI array processors, AI algorithms, machine learning, deep learning networks, neural architectural search, and compressive privacy. He was a founding member of several Technical Committees of the IEEE Signal Processing Society. He was elected to Fellow of IEEE in 1988 and served as a Member of the Board of Governors of the IEEE Signal Processing Society (1989-1991). He was a recipient of IEEE Signal Processing Society's Technical Achievement Award for the contributions on "parallel processing and neural network algorithms for signal processing" (1992); a Distinguished Lecturer of IEEE Signal Processing Society (1994); a recipient of IEEE Signal Processing Society's Best Paper Award (1996); IEEE Third Millennium Medal (2000), and CIE-USA’s AAEOY Distinguished Achievement Award (2023). Since 1990, he has been the Editor-In-Chief of the Journal of VLSI Signal Processing Systems. He has authored and co-authored more than 500 technical publications and numerous textbooks including ``VLSI Array Processors'', Prentice-Hall (1988); ``Digital Neural Networks'', Prentice-Hall (1994) ; ``Principal Component Neural Networks'', John-Wiley (1996); ``Biometric Authentication: A Machine Learning Approach'', Prentice-Hall (2004); and ``Kernel Methods and Machine Learning”, Cambridge University Press (2014).

Abstract: Over the past decade, cyber-physical systems (CPS) have transformed our daily lives by seamlessly integrating computational intelligence into physical world. The synergy between digital twin and big data analytics emerges as a pivotal force that drives smart network operations and maintenance for CPS. Network digital twin provides a dynamic, high-fidelity virtual model of physical networks, while big data techniques offer powerful tools to efficiently process large volumes of network data. This talk will delve into the transformative potential of this synergy in the design, operation, and optimization of future networking systems. Our vision focuses on harnessing the collective power of digital twin and big data in network analytics and modelling, aiming to facilitate real-time system monitoring, proactive troubleshooting, automated optimization, and intelligent decision-making. This talk will further present cutting-edge methodologies in network big data modelling, real-time incremental data analytics, and cost-effective distributed computing platforms, to achieve accurate and timely anomaly detection and predictive analysis for network digital twin, towards more intelligent, secure, reliable and responsive future CPS.

Bio: Professor Geyong Min is a Chair in High Performance Computing and Networking in the Department of Computer Science at the University of Exeter, UK. His research interests include Computer Networks, Cloud and Edge Computing, Mobile and Ubiquitous Computing, Systems Modelling and Performance Engineering. His recent research has been supported by European Horizon-2020, UK EPSRC, Royal Society, Royal Academy of Engineering, and industrial partners. He has published more than 200 research papers in leading international journals including IEEE/ACM Transactions on Networking, IEEE Journal on Selected Areas in Communications, IEEE Transactions on Computers, IEEE Transactions on Parallel and Distributed Systems, and IEEE Transactions on Wireless Communications, and at reputable international conferences, such as SIGCOMM-IMC, INFOCOM, and ICDCS. He is an Associated Editor of several international journals, e.g., IEEE Transactions on Computers, and IEEE Transactions on Cloud Computing. He served as the General Chair or Program Chair of a number of international conferences in the area of Information and Communications Technologies

Abstract: Artificial intelligence is developing rapidly today and is reshaping all walks of life. Yuanqi Data Asset Platform is a flagship product in the data field of Gientech. Its goal is to help enterprises realize data collection, processing, processing, analysis and application. It is an end-to-end data analysis tool and has achieved great application results in the financial industry. In the AI era, Yuanqi Data Asset Platform has been further iterated and upgraded by leveraging the capabilities of AI. Driven by the real needs of customers, great progress has been made in code generation, intelligent reporting and digital twins, which has greatly improved development efficiency and customer experience. This is an important progress of AI in the data field of the financial industry.

Bio: Mr. Du Xiaozheng graduated from the Beijing University of Aeronautics and Astronautics with a major in systems engineering and received an EMBA from Peking University. He is the deputy director of GienTech Research Institute and the general manager of the business analysis division. He is also a standing member of the CCF Data Governance Development Committee, an executive member of the CCF Big Data Committee, a member of the CCF CTO Club, and a director of DAMA Greater China. He has worked for many large companies such as Alibaba Cloud, Lenovo Group, Teradata, and Southeast Rongtong, focusing on the content research of industry big data platform construction, data governance system construction, data middle platform products, and data ecosystem application construction. In the financial industry, he has participated in constructing big data platforms, data governance, and applications for many large state-owned banks, joint-stock banks, and small and medium-sized banks. During his time at Alibaba Cloud, he participated in many large digital government and central enterprise data projects and has a deep understanding of government and enterprise data governance. Du Xiaozheng led Gientech big data solution team and won the first place in the IDC big data solution market for seven consecutive years. The Yuanqi Data Asset Platform he designed has been appraised by the Electronics Society, reaching the leading level in China.