Speakers of EITCE 2023

Prof. Yingxu Wang, University of Calgary, Canada (IEEE Fellow, AAIA Fellow)(h-index 67)

Dr. Yingxu Wang is professor of cognitive systems, brain science, software science, and intelligent mathematics. He is the founding President of International Institute of Cognitive Informatics and Cognitive Computing (I2CICC). He is FIEEE, FBCS, FI2CICC, FAAIA, and FWIF. He has held visiting professor positions at Univ. of Oxford (1995, 2018-2022), Stanford Univ. (2008, 2016), UC Berkeley (2008), MIT (2012), and a distinguished visiting professor at Tsinghua Univ. (2019-2022). He received a PhD in Computer Science from Nottingham Trent University, UK, in 1998 and has been a full professor since 1994. He is the founder and steering committee chair of IEEE Int'l Conference Series on Cognitive Informatics and Cognitive Computing (ICCI*CC) since 2002. He is founding Editor-in-Chiefs and Associate Editors of 10+ Int'l Journals and IEEE Transactions. He is Chair of IEEE SMCS TC-BCS on Brain-inspired Cognitive Systems, and Co-Chair of IEEE CS TC-CLS on Computational Life Science.,His basic research has spanned across contemporary scientific disciplines of intelligence, mathematics, knowledge, robotics, computer, information, brain, cognition, software, data, systems, cybernetics, neurology, and linguistics. He has published 600+ peer reviewed papers and 38+ books/proceedings. He has presented 65+ invited keynote speeches in international conferences. He has served as honorary, general, and program chairs for 40+ international conferences. He has led 10+ international, European, and Canadian research projects as PI. He is recognized by Google Scholar as world top 1 in Software Science, top 1 in Cognitive Robots, top 8 in Autonomous Systems, top 2 in Cognitive Computing, and top 1 in Knowledge Science with an h-index 67. He is recognized by ResearchGate as among the world's top 1.0% scholars in general and in several contemporary fields encompassing artificial intelligence, autonomous systems, theoretical computer science, engineering mathematics, software engineering, cognitive science, information science, and computational linguistics. He has published formal proofs for two of the world's top ten hardest mathematical problems known as the Goldbach conjecture and Twin-Prime conjecture in 2022.

Distinguished Professor Philippe Fournier-Viger, Shenzhen University, China (Youth 1000 National Talent in China)

Philippe Fournier-Viger (Ph.D) is a Canadian researcher, distinguished professor at Shenzhen University (China). Five years after completing his Ph.D., he came to China in 2015 and became full professor after receiving a national talent title from the National Science Foundation of China. He has published more than 375 research papers related to data mining algorithms for complex data (sequences, graphs), intelligent systems and applications, which have received more than 12,000 citations (cf. Google Scholar). He is the founder of the popular SPMF data mining library, offering more than 250 algorithms to find patterns in data, cited in more than 1,000 research papers. He is former associate edito-in-chief of the Applied Intelligence journal and has been keynote speaker for over 25 international conferences and co-edited four books for Springer. He is a co-founder of the UDML, PMDB and MLiSE series of workshops held at the ICDM, PKDD, DASFAA, and KDD conferences.

Title: Advances for the Automatic discovery of Interesting and useful patterns in data

Abstract:Intelligent systems and electronic tools can play an important role in various domains such as for factory automation, e-business, and manufacturing. To build intelligent systems and tools, high-quality data is generally required. Moreover, these systems need to process complex data and can yield large amounts of data such usage logs, images, videos, and data collected from industrial sensors. Managing this data to gain insights and improve these systems is thus a key challenge. It is also desirable to be able to extract information or models from data that are easily understandable by humans. Based on these objectives, this talk will discuss the use of data mining algorithms for discovering interesting and useful patterns in data generated from intelligent systems and other applications.

The talk will first briefly review early study on designing algorithms for identifying frequent patterns. Then, an overview of recent challenges and advances will be presented to identify other types of interesting patterns in more complex data such as sequences and dynamic graphs. Topics that will be discussed include high utility patterns, locally interesting patterns, and periodic patterns. Lastly, the SPMF open-source software will be mentioned and opportunities related to the combination of pattern mining algorithms with traditional artificial intelligence for intelligent systems will be discussed.

Prof. Yuanlong Yu (Dean of the College), Fuzhou University, China

于元隆教授，福州大学计算机与大数据学院院长

Dr. Yuanlong Yu is a Professor of Computer Science at Fuzhou University. He is the Director of China Association of Command and Control, and Vice Director-general of Fujian Association of Artificial Intelligence. Since 2021, he has been the Dean of College of Computer and Data Science at Fuzhou University. Dr Yu’s research area are computer vision, machine learning and cognitive robotics. He has been granted 4 projects from National Natural Science Foundation of China. He has published more than 70 papers, including 10 CCF-A rank conference and journal publications. He has respectively obtained Fujian Province 2nd Prize of Progress in Science and Technology in 2019 and 2021, and also obtained Fujian Province Grand Prize of Education and Teaching Achievement in 2022. 

Title：Exploring the Generalizability of Computational Models for Visual Perception

Abstract: Traditional machine learning algorithms have achieved great success on various vision tasks under identical and independent (IID) assumption. However, most cases in the real world are incompatible with IID assumption, i.e., the probability distributions of training data and testing data are different. It leads to tremendous performance degradation due to domain shift. Domain generalization is a promising way to deal with domain shift in the unknown task domain is necessary. This talk gives some of our recent work about domain generalization. First, a dynamic domain generalization framework by integrating dynamic networks is presented. Furthermore, a semi-supervised domain generalization algorithm is proposed for some unlabeled source domain. Some experimental results have also been shown in the talk to evaluate our proposed algorithms. 

Prof. Yue Huang, Xiamen University, China

黄悦教授，厦门大学

Yue Huang, is a full professor at the School of Information, Xiamen University. She also served as the vice director of the Department of Information and Communication Engineering. She received a doctorate in engineering from Tsinghua University and a bachelor's degree in engineering from Xiamen University. Her research interests include machine learning theory and its application in biomedical image and radar signal analysis. She is the primary investigator many National Natural Science Foundation of China. He won the China Computer Federation-Tencent Rhino-Bird Fund twice (11% selection rate) and won an Outstanding Achievement Award. As one of the main contributors, she won the third prize of the Wu Wenjun Artificial Intelligence Natural Science Award in 2019, the third prize of the Fujian Provincial Science and Technology Progress Award in 2017 and the third prize of the Fujian Provincial Natural Science Award in 2020. She has published many papers in important journal conferences including IEEE Transactions, CVPR/ICCV/ECCV/ICASSP and other fields. She served as the associate editor of IEEE Trans. Neural Networks and Learning Systems. She is the organizing committee member of the IEEE MMSP2015, IEEE ISIMCT 2021, and IEEE ICCC 2021.

Title: Domain Adaptive object detection in the open world

Abstract: Domain Adaptive Object Detection (DAOD) focuses on improving the generalization ability of object detectors via knowledge transfer. Recent advances in DAOD strive to change the emphasis of the adaptation process from global to local in virtue of fine-grained feature alignment methods. However, both the global and local alignment approaches fail to capture the topological relations among different foreground objects as the explicit dependencies and interactions between and within domains are neglected. In this case, only seeking one-vs-one alignment does not necessarily ensure the precise knowledge transfer. Moreover, conventional alignment-based approaches may be vulnerable to catastrophic overfitting regarding those less transferable regions (e.g. backgrounds) due to the accumulation of inaccurate localization results in the target domain. To remedy these issues, we first formulate DAOD as an open-set domain adaptation problem, in which the foregrounds and backgrounds are seen as the “known classes” and “unknown class” respectively. Accordingly, we propose a new and general framework for DAOD, named Foreground-aware Graph-based Relational Reasoning (FGRR), which incorporates graph structures into the detection pipeline to explicitly model the intra- and inter-domain foreground object relations on both pixel and semantic spaces, thereby endowing the DAOD model with the capability of relational reasoning beyond the popular alignment-based paradigm. FGRR first identifies the foreground pixels and regions by searching reliable correspondence and cross-domain similarity regularization respectively. The inter-domain visual and semantic correlations are hierarchically modeled via bipartite graph structures, and the intra-domain relations are encoded via graph attention mechanisms. Through message-passing, each node aggregates semantic and contextual information from the same and opposite domain to substantially enhance its expressive power. Empirical results demonstrate that the proposed FGRR exceeds the state-of-the-art performance on four DAOD benchmarks.

Assoc. Prof. Xiufeng Liu (senior researcher), Technical University of Denmark (DTU), Denmark

刘秀峰副教授，丹麦技术大学

Dr. Xiufeng Liu is a senior researcher (Associate Professor level) at the Department of Technology, Management and Economics at the Technical University of Denmark (DTU). He specializes in smart meter data analysis, energy informatics, Big Data management and ETL. He has authored or co-authored more than 100 papers in prestigious energy and database journals and conferences, including Applied Energy, Energy, IEEE TSG, TII, TNNLS, EDBT, ICDE and TODS. He received the best paper award in EDBT 2015. Dr. Liu obtained his PhD in Computer Science from Aalborg University, Denmark, in 2012. He then worked as a postdoctoral researcher in the data systems group at the University of Waterloo, Canada, and as a research scientist at IBM Toronto Research Center in 2013-2014. He joined DTU as a faculty member in 2015. He has been involved in, led or co-led over ten research projects funded by Danish Innovation Fund, European Horizon 2020, ERA-NET and Marie Curie programs.

Title: Data to Energy: How to Manage and Analyze Energy Information

Abstract:Energy data is a rich source of insights for enhancing energy efficiency, reliability, security, and sustainability. However, energy data also poses many challenges, such as data integration, data analysis, data visualization, data prediction, and data decision making. In this talk, a data intelligence approach for smart energy systems will be introduced, which involves modeling, collecting, cleansing, integrating, understanding, predicting, and acting on energy data. Some applications of data intelligence for smart energy systems will be showcased, such as smart meter data analytics system, spatiotemporal visual analysis tool for energy consumption patterns, and spatiotemporal energy demand prediction. Moreover, the role of visual analytics in understanding energy consumption patterns from spatial, temporal, and spatiotemporal dimensions will be highlighted. A novel potential flow based method to model energy demand shift patterns across spatial and temporal dimensions will be demonstrated. With this method and tool, users can interactively discover the spatiotemporal variability of energy demand, customer groups, and demand shift patterns across different geographical areas and time horizons. The talk will conclude with some future directions and opportunities for advancing the field of data to energy.

Invited  Speaker：

Prof. Xiaobo Qu, Xiamen University, China

屈小波教授，厦门大学

Xiaobo Qu received his Ph.D. degree in communication engineering from Xiamen University, Xiamen, China, where he is a professor and the vice director of the Department of Electronic Science, the director of the Biomedical Intelligent Cloud Research and Development Center, and the vice director of the Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance. He received the Distinguished Young Scholar Award from the Natural Science Foundation of Fujian Province of China (2018), He Yici Chair Professor Award (2018) from Xiamen University, Excellent Young Scientists Fund Award from the National Natural Science Foundation of China (2021), First Prize in Natural Science Award, and Yunsheng Youth Science and Technology Award of Fujian Province of China (2022), and Liu Xiao Mathematical and Electronic Science Award from Xiamen University (2023). He is an associate editor of IEEE Transactions on Computational Imaging, a senior editor of BMC Medical Imaging, an associate editor of Magnetic Resonance Letters, etc. His research interests include magnetic resonance imaging and spectroscopy, computational imaging, machine learning, artificial intelligence, and cloud computing. Qu is a senior member of IEEE and a member of ISMRM.

Title: Physics-informed Artificial Intelligence for Biomedical Magnetic Resonance Imaging

Abstract:Magnetic resonance imaging (MRI) is an indispensable non-invasive and non- radioactive diagnosis tool in biomedical imaging. One of its fundamental problem is the relatively slow imaging speed. To accelerate the imaging, state-of-the-art artificial intelligence (AI) methods try to learn nonlinear mappings from the undersampled data to the artifacts-free images and/or spectra. However, due to the lack of high-quality or even unavailable training data, AI encounters bottleneck in challenging scenarios (e.g., motion artifacts, low signal-to-noise ratio, and novel imaging sequences). How to break through this data bottleneck, achieve interpretable and trusted AI are frontier issues. This lecture focuses on a new paradigm for computational MRI, the physics-informed AI (PIAI) that can learn physical rule of MRI signal generation. Starting from the MRI Bloch differential equations, two main PIAI forms, implicit learning with physics-driven synthetic data and explicit learning with differential equations will be discussed. PIAI is expected to provide physics interpretable and reliable AI imaging, boosting the next generation development of MRI equipment and precision medicine.