Prof. Yinglin Wang
Professor at School of Information Management and Engineering,
Shanghai University of Finance and Economics, Shanghai, China
Investigating and Mitigating the Bias in Textual Datasets for Causal Learning
Supervised learning algorithms recklessly absorbing all the correlations found in training data is statistically correct but might have missed the point. Although those models have promising results, they can still be problematic for several reasons. Firstly, the generation of the models could be hurt when it meets the test set which has different distribution from the training set. Additionally, the models could capture and amplify the social bias when the correlation involves the demographic attributes. In addition, the models which absorb the spurious correlations suffer from the trust issues while the prediction need to be explained to the human. This talk presents some recent studies, in which we explore the bias in the datasets for NLP tasks such as commonsense causal reasoning and sentiment analysis from text, which need to capture causal features, not just correlations. We introduce the methods to alleviate the problem from the perspective of data and model, and discuss the future research directions.
Wang Yinglin is a full professor in the School of Information Management and Engineering, Shanghai University of Finance and Economics. He was Previously a full Professor in the Department of Computer Science at the Shanghai Jiao Tong University. Prior to that, he was a postdoc at Shanghai Jiao Tong University. He completed his Ph.D. in pattern recognition and intelligent control from Nanjing University of Science and Technology in 1998. In 2001 and 2005, he visited the University of Hong Kong and Stanford University for cooperative research. His current research interests are broadly in data mining, knowledge management, and software engineering. His research work has been published at top NLP conferences including ACL, EMNLP and well-known journals such as KBS and IPM. He has taken charge of many research projects sponsored by NSFC etc. He is a recipient of two Shanghai science and technology progress awards.
Prof. Dr. Ali Selamat
Fellow, Media & Games Center of Excellence, Universiti Teknologi Malaysia &
Dean, Malaysia Japan International Institute of Technology (MJIIT), Universiti Teknologi Malaysia
Fellow, Academy Professor Malaysia (APM)
Recent Research on Detection of Vulnerable Plaque in Virtual Histology Intravascular Ultrasound Images
Virtual Histology Intravascular Ultrasound (VH-IVUS) image has been developed as a clinically available tools to diagnosis the thin cap fibroatheroma plaque that is responsible for coronary artery disease. Segmentation is considered as the essential step for plaque characterization; however, this step is time-consuming procedure and depends on the observer's capability. VH-IVUS image, especially at the edges of each plaque component, may not involves the unique intensities. Therefore, segmentation of the overlapped tissue can be difficult. In this study, hybrid models have been proposed to accurately segment the VH-IVUS image. Therefore, semi-supervised models of hybrid Fuzzy C-Means (FCM) with k-nearest neighbor (kNN), minimum Euclidian distance, and Support Vector Machine (SVM) are adapted to improve the segmentation results. Proposed approaches is applied for 589 in-vivo images of 10 patients. The accurate classification result of the semi-supervised models is obtained using the silhouette validity index.
Thin cap fibroatheroma (TCFA), VH-IVUS segmentation, semi-supervised model, clustering optimization, support vector machines, machine learning
Ali Selamat has received a B.Sc. (Hons.) in IT from Teesside University, U.K. and M.Sc. in Distributed Multimedia Interactive Systems from Lancaster University, U.K. in 1997 and 1998, respectively. He has received a Dr. Eng. degree from Osaka Prefecture University, Japan, in 2003.
Ali Selamat is currently a professor at the School of Computing, Faculty of Engineering, Universiti Teknologi Malaysia (UTM). He is presently serving as a Dean of Malaysia Japan International Institute of Technology (MJIIT), Universiti Teknologi Malaysia.Before that, he was a Chief Information Officer (CIO) and a Director of Communication and Information Technology Director, UTM. He is currently elected as a Chair of IEEE Computer Society, Malaysia Section under the Institute of Electrical and Electronics Engineers (IEEE), USA, and a Malaysia Engineering Deans Council member. He was previously assuming the position of research Dean on Knowledge Economy Research Alliance, UTM. He is currently elected as a fellow under Academy Professor Malaysia and a research fellow at Magicx - Media and Games Center of Excellence, Universiti Teknologi Malaysia.
He was a principal consultant of Big Data Analytics, Ministry of Higher Education, Malaysia 201, and currently, a member of Malaysia Artificial Intelligence Roadmaps (2020-2021) and a keynote speaker in many international conferences. He was a visiting professor at Kuwait University and few other universities in Japan, Saudi Arabia, and Indonesia. Currently, he is a visiting professor at Hradec-Kralove University, Czech Republic, and Kagoshima Institute of Technology, Japan.
He was serving as the Editorial Boards of International Journal of Knowledge-Based Systems Elsevier, Netherlands, and currently an associate editor for International Journal of Artificial Intelligence and Machine Learning (IJAIML), IGI Global and Journal of Service Oriented Computing and Application (SOCA), Springer.
He is also the editorial members of International Journal of Information and Database Systems (IJIIDS) under Inderscience Publications, Switzerland, and Vietnam Journal of Computer Science under Springer Publications. He is the Program co-chair of IEA/AIE 2021: The 34th International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems in Kuala Lumpur, Malaysia.
His research interests include data analytics, digital transformations, knowledge management in higher educations, key performance indicators, cloud-based software engineering, software agents, information retrievals, pattern recognition, genetic algorithms, neural networks, and soft-computing.