CNCC2024
Brief introduction of the forum:
A large spatiotemporal model of the network
Time: 13:30-17:30 on October 25
Location: 3rd Floor, Summer Garden
Note: If there is any change, please refer to the final information on the official website (https://ccf.org.cn/cncc2024).
Driven by big data and artificial intelligence technologies, cyber spatiotemporal models have become an indispensable tool for analyzing and predicting complex network systems, and are widely used in many key fields such as smart cities, intelligent transportation, digital energy, cybersecurity, intelligent O&M, social networks, and AI for Science.
However, the traditional spatiotemporal data processing technology still faces challenges such as low efficiency, insufficient feature mining, and difficult algorithm migration, which leads to many problems in the existing spatiotemporal models in dealing with multi-domain, multi-task (such as prediction, detection, classification, inference, etc.), multi-subdivision modalities (including unvariate, multivariate, graph structure, table structure, etc.), and insufficient data and samples.
In order to address these challenges, the Foundation Model of NETS (Network of Time Series) is expected to achieve a major breakthrough in the research of the Foundation Model of NETS (Network of Time Series). This seminar will deeply discuss the research issues, data foundations, methodologies and practical techniques of network spatio-temporal large models, covering the theory and practice of dynamic graph large models, the intelligent management and analysis methods of spatio-temporal big data, and the construction and application of spatio-temporal basic models. The speakers will comprehensively discuss the generality, automation, robustness, explainability and lightweight of spatiotemporal data analysis and modeling methods, from the efficient collection and intelligent computing of spatiotemporal big data, to data governance, analysis and decision-making, to the adaptive and generalized modeling of spatiotemporal basic models, as well as the collaboration of large and small models and the orchestration and implementation of multi-agents.
Forum Agenda
| order | topic | Keynote speaker | unit |
| 1 | Dynamic spatiotemporal graph neural networks, large models and applications | Zhu Wenwu | Tsinghua University |
| 2 | Spatiotemporal big data intelligent computing | Gao Yunjun | Zhejiang University |
| 3 | Fine-grained network monitoring, time series data collection and transmission, and low overhead | Xie Kun | Hunan University |
| 4 | Time-series and spatiotemporal data-driven decision intelligence | Yang Bin | East China Normal University |
| 5 | Time-series intelligence in cyberspace | Pei Dan | Tsinghua University |
| 6 | Panel link | Zhu Wenwu | Tsinghua University |
| Gao Yunjun | Zhejiang University | ||
| Xie Kun | Hunan University | ||
| Yang Bin | East China Normal University | ||
| Li Yong | Tsinghua University | ||
| Pei Dan | Tsinghua University |
Introduction of the chairman and guests of the forum
Chair of the Forum
Sun Yongqian
Associate Professor, School of Software, Nankai University
Associate professor and doctoral supervisor of the School of Software, Nankai University. He graduated from Tsinghua University with a Ph.D. in Computer Science, is a senior member of CCF, and an executive member of CCF's Internet, Software Engineering, Service Computing, and Architecture committees. He has published more than 50 papers in international conferences or journals recommended by CCF, such as JSAC, WWW, ASE, FSE, TC, TSC, TOSEM, etc., in the field of intelligent operation and maintenance (AIOps), research on fault identification and diagnosis based on multimodal data, and published more than 50 papers in international conferences or journals recommended by CCF, such as JSAC, WWW, ASE, FSE, TC, TSC, TOSEM, etc. He has presided over 1 National Natural Youth Project, 1 Tianjin Youth Fund, and more than 10 school-enterprise joint research projects with Huawei, Alibaba, Tencent, ByteDance, Kuaishou, etc. Won the first prize of scientific and technological progress of the Chinese Institute of Electronics.
Forum Speaker
Zhu Wenwu
He is a fellow of CCF, deputy director of the Big Data Committee, a professor of Tsinghua University, and the deputy director of the National Research Center for Information Science and Technology
Professor of the Department of Computer Science of Tsinghua University, Deputy Director of the National Research Center for Information Science and Technology, Deputy Director of the National Engineering Laboratory for Big Data Algorithms and Analysis, Director of the Big Data Intelligence Research Center of the Institute of Artificial Intelligence of Tsinghua University, Chief Scientist of the National 973 Project, and Leader of Major Projects of the National Foundation of China. He has won the second prize of the National Natural Science Award three times. He is currently the editor-in-chief of IEEE Transactions on Circuits and Systems for Video Technology, and has served as the chairman of the IEEE Transactions on Multimedia Steering Committee and the editor-in-chief of IEEE Transactions on Multimedia. Winner of the 2023 ACM SIGMM Technical Achievement Award, and the 2024 IEEE Circuits and Systems Society Charles A. Desoer Technical Achievement Award. ACM Fellow, AAAS Fellow, IEEE Fellow, SPIE Fellow, Foreign Member of the European Academy of Sciences.
Title: Dynamic Spatiotemporal Graph Neural Networks, Large Models and Their Applications
Abstract: The report will introduce the basic concepts, challenges, research progress and future research directions of dynamic graph machine learning. Firstly, dynamic graph embedding representation learning and dynamic graph neural networks are introduced. Then, the new research progress of dynamic graph machine learning is introduced: 1) spatiotemporal graph neural networks in dynamic open environments, including automatic graph machine learning and out-of-distribution generalized dynamic graph machine learning, that is, mining time-varying and invariant patterns in the case of non-independent and identical distribution of training and test data to make the model adaptive and generalized; 2) Dynamic graph large model, that is, the dynamic graph task is unified to adapt to different scenarios and different needs through large models. Finally, the future research directions are introduced.
Gao Yunjun
Qiushi Distinguished Professor and Vice Dean of the School of Software at Zhejiang University, Vice Chairman of ACM China SIGPATIAL, and Director of Zhejiang Provincial Key Laboratory of Big Data Intelligent Computing
Qiushi Distinguished Professor and Doctoral Supervisor of Zhejiang University, Deputy Dean of the School of Software, Director of Zhejiang Provincial Key Laboratory of Big Data Intelligent Computing, National Outstanding Young Scholar and National Excellent Youth Fund Winner. His research interests include databases, big data management and analysis, and the integration of databases and artificial intelligence. He has published more than 150 CCF Class A papers, published 4 monographs, has more than 20 authorized patents, and won 6 best/excellent papers in ICDE and other conferences, and 3 provincial and ministerial special prizes or first prizes. He has served as the vice chairman of ACM China SIGPATIAL, and has served as an editorial board member in journals such as TKDE, and as a local chair or member of the program committee at meetings such as VLDB.
Title: Intelligent Computing of Spatio-temporal Big Data
Abstract: Spatiotemporal big data intelligent computing is an important support for major national needs such as digital economy and smart cities. Traditional spatiotemporal data processing technology still faces challenges such as low computing efficiency, shallow feature mining, and difficult algorithm migration. Based on database and big data technology, combined with artificial intelligence methods, the speaker carried out research on intelligent management and analysis of spatiotemporal big data, including embedded representation, spatiotemporal compression, intelligent query, deep mining and multimodal spatiotemporal fusion. This report focuses on the intelligent computing of spatiotemporal big data, first summarizes the speaker's thinking on the intelligent management and analysis of spatiotemporal big data, and then introduces the team's progress and prospects in this field in recent years.
Xie Kun
He is a second-level professor of Hunan University and the director of the Key Laboratory of Supercomputing and Artificial Intelligence Convergence Computing of the Ministry of Education
He is a second-level professor of Hunan University, a doctoral supervisor, a winner of the National Fund for Distinguished Young Scholars, a winner of the Hunan Provincial Fund for Outstanding Young Scholars, and a pacesetter of "Women Meritorious Service" in Changsha City. His research interests include intelligent operation and maintenance and security of network systems, computer networks, big data and artificial intelligence. In the past five years, he has published more than 60 papers in international conferences such as SIGMOD and INFOCOM and IEEE journals, and the research results have been applied to the actual network platform. He is the director of the Key Laboratory of Supercomputing and Artificial Intelligence Convergence Computing of the Ministry of Education.
Title: Low-overhead collection and transmission of fine-grained network monitoring time series data
Report Summary: With the development of AI, there is a growing interest in applying AI to understand and develop network systems for network operation and maintenance, which increases the demand for network measurement data. However, traditional minute-level network KPI monitoring cannot capture microbursts (sudden drops), and even with powerful AI tools, it is difficult to diagnose and locate network system problems and faults. When the granularity of network monitoring changes from the traditional minute level to the second level and millisecond level, the measurement collection, transmission, and storage overhead increases thousands of times and thousands of times. This report introduces the challenges and solutions for fine-grained KPI time series data acquisition and transmission.
Yang Bin
Chair Professor, East China Normal University
Chair Professor of School of Data Science and Engineering, East China Normal University, National Leading Talent, Ph.D. Supervisor. He was a distinguished scientist and tenured professor at the Faculty of Engineering and Technology, Aalborg University, Denmark, and was funded by the Sapere Aude scientist (Denmark National Talent Program) from the Denmark Independent Research Foundation. His research interests include decision intelligence, time series analysis, spatiotemporal data analysis, data management and analysis, AI for Science, etc. He has published more than 90 papers in important international conferences and journals, and the results have been widely used in relevant enterprises and government departments in China, Denmark, Germany, Netherlands, Greece and Cyprus.
Title: Time series and spatiotemporal data-driven decision intelligence
Abstract: Time series and spatiotemporal data-driven decision intelligence aims to efficiently govern and intelligently analyze multi-source heterogeneous multimodal time series and spatio-temporal data with the two dimensions of time and space as the traction, so as to effectively support data-driven decision-making and empower applications such as smart cities, intelligent transportation, digital energy, intelligent O&M, and AI for Science. Focusing on the research paradigm of "data-governance-analysis-decision-making", the speaker will introduce the multimodal spatiotemporal data foundation and data governance methods of time series, elaborate on the universality, automation, robustness, interpretability and lightweight of data analysis methods, discuss the corresponding benchmarking, and finally introduce different types of data-driven decision-making strategies.
Pei Dan
He is an associate professor in the Department of Computer Science, Tsinghua University
He is a tenured associate professor and doctoral supervisor in the Department of Computer Science, Tsinghua University. His main research interests are machine learning-based intelligent operation and maintenance (AIOps) and network time series intelligence. He has published more than 200 academic papers and more than 30 patents in the field of intelligent operation and maintenance, and has been cited more than 10,000 times by Google Scholar. Won the first prize of scientific and technological progress of the Chinese Institute of Electronics. He is the founder of the CCF International AIOps Challenge and the initiator of the CCF OpenAIOps community, which has been successfully held for seven times. He serves on the editorial board of IEEE/ACM Transactions on Networking, the flagship journal in the field of computer networks, and has served as the chairman of the technical program committee of ICNP 2022, the flagship conference in the field of computer networks of IEEE.
Title: Time Series Intelligence in Cyberspace
Abstract: Time series intelligence is an indispensable ability for general artificial intelligence to perceive, recognize the real world, and make decisions. It has the characteristics of multi-domain, multi-task (prediction, anomaly detection, classification, causal inference, etc.), and multi-subdivision modality (unary, multivariate, graph structure, table structure, etc.), which brings great challenges to traditional temporal machine learning methods, resulting in its inability to plug and play in reality, and difficulty in implementation. This report will share the recent research progress and application of cyberspace time series intelligence "small models" and basic models, and discuss how to train time-series basic models for general-purpose, plug-and-play, low-cost inference.
About CNCC2024
CNCC2024 will be held on October 24-26 in Hengdian Town, Dongyang City, Zhejiang Province, with the theme of "Developing New Quality Productivity, Computing Leads the Future". The three-day conference included 18 invited reports, 3 conference forums, 138 thematic forums, 34 thematic activities and more than 100 exhibitions. More than 800 speakers, including Turing Award winners, academicians of the Chinese Academy of Sciences and the Chinese Academy of Sciences, top scholars at home and abroad, and well-known entrepreneurs, looked forward to cutting-edge trends and shared their innovative achievements. More than 10,000 people are expected to attend.