2023 3rd International Conference on Digital Society and Intelligent Systems (DSInS 2023)

Keynote Speakers

Keynote Speakers

刘斌Bing Liu.png

Prof. Bing Liu

AAAI/ACM/IEEE Fellow, University of Illinois at Chicago (UIC), United States

Title: Unifying Continual Learning, OOD Detection and Open World Learning

Abstract: Continual learning aims to incrementally learn a sequence of tasks. A challenging setting of CL is class incremental learning (CIL). This talk first presents a theoretical study on how to solve the CIL problem. The key result is that the necessary and sufficient conditions for good CIL are good within-task prediction and good out-of-distribution (OOD) detection. The theory thus unifies CIL and OOD detection. Based on the theory, new CIL methods have been designed which significantly outperform existing CIL baselines. Interestingly, the theory is also applicable to open world learning (OWL) because OWL needs to detect unknowns or novelties (which is OOD detection) after model deployment and incrementally learn the detected knowns or new objects on the fly (which is CIL). OWL is by nature continual and has the goal of building autonomous learning agents that can learn by themselves in a self-motivated and self-initiated manner rather than by being retrained offline periodically as ordered by human engineers.


Bing Liu is a distinguished professor at the University of Illinois Chicago. He received his Ph.D. in Artificial Intelligence (AI) from the University of Edinburgh. His current research interests include continual/lifelong learning, lifelong learning dialogue systems, sentiment analysis, machine learning and natural language processing. He has published extensively in prestigious conferences and journals and authored four books: one about lifelong machine learning, two about sentiment analysis, and one about Web mining. Three of his papers have received the Test-of-Time awards, and another one received Test-of-Time honorable mention. Some of his works have also been widely reported in popular and technology press internationally. He served as the Chair of ACM SIGKDD from 2013-2017 and as program chair of many leading data mining conferences. He is also the winner of 2018 ACM SIGKDD Innovation Award, and is a Fellow of ACM, AAAI, and IEEE.


Prof.Bo Peng

Southwest Jiaotong University, China

Title: Robust Edge and Boundary Segmentation in Urban Pavement Defects Detection

Abstract: Edge and boundary segmentation is a basic task in the field of computer vision and can be used as a basic operation in many complex tasks. However, there are still many challenges to overcome in urban pavement defects detection owing to the complexity of the image background and the existence of edge noise. As a critical step in pavement defects detection,recent deep convolutional neural network (DCNN) models have been widely used, and its performance has been greatly improved compared to traditional methods. Due to the crisp-edge prediction problem, the background pixels near the edge of the object are easily misclassified, results in a relatively thick edge in the final prediction. To address the challenge of imbalanced data resulting from the prevalence of non-crack pixels, we seek to improve the quality of pavement crack segmentation, particularly for thick and tiny cracks. Moreover, some bio-inspired algorithms are studied to solve the edge connection and defects detection problems under challenging environments. They show promising results for deploying the algorithms in Edge devices.

Experience: Bo Peng is a professor in the School of Computing and Artificial Intelligence, Southwest Jiaotong University, Chengdu, China. She received the M.S. degree from the Department of Computer Science, Western University and the Ph.D. degree from the Department of Computing, The Hong Kong Polytechnic University. Her research interests include deep learning, semantic image segmentation, probabilistic graph models, industrial defects detection and medical image segmentation. She has published more than 70 papers in journals and conferences, including TPAMI, TIP, PR, ECCV, ACCV, BMVC and so on. More than 20 patents have been approved/applied. She has won the best poster papers of ISKE, ACM Chengdu "New Star Award". 



Asia University, Taichung Taiwan, China

Title: Combing Multi-Biometric Recognition Techniques to Smart Cities


The dataset [1, 2, 4, 5] demonstrates the wide array of applications that have been developed through Big Data processing, including Weather forecasting, Stock market analysis, Traffic control, Energy conservation, and Surveillance management. The rapid generation of a massive volume of messages on the internet has given rise to several challenges, such as open access, security, privacy concerns, and uncertainties.

The desire for the convenience of living in a smart city is well-understood [3], but it brings with it a significant dilemma. The more advanced and intelligent cities become, the larger the volume of data they generate. The task of handling and analyzing such enormous data sets in the field of computer engineering is an intricate and demanding one.

Instead of traditional methods like keys or tickets for access control, a more effective approach is using biometrics [6-12]. Biometrics involves recognizing individuals based on their unique physical traits, like fingerprints, iris, veins, heartbeats, face, and voice. It is a reliable system that uses these distinctive characteristics to verify and identify people. This implies that there is no need to worry about losing keys or tickets when accessing locations or transportation.

Deep learning is a machine learning technique inspired by the human brain. It uses multi-layer neural networks to solve complex problems. These networks automatically learn from a lot of data to perform tasks like image recognition, language understanding, stock market predictions, self-driving cars, medical diagnosis, and more. The key is that it keeps getting better by learning from data.

In smart cities everything should be smart and it is quite user friendly to use fingerprints, iris, veins, heart beats, face and voice which are owned by user itself for authentication control. The use of deep learning techniques in biometric recognition for access control is increasingly significant. To enhance security, the adoption of fusion techniques to combine multiple biometric recognition methods is gaining attention.

In this talk, we will introduce Big Data, Smart Cities and Multi-Biometric Recognition fusing technologies and point out the problems we face and need to be solved.


SHI-JINN HORNG received the B.S. degree in electronics engineering from the National Taiwan Institute of Technology, the MS degree in information engineering from the National Central University, and the Ph.D. degree in computer science from the National Tsing Hua University in 1980, 1984, and 1989, respectively. He was a professor and dean of the College of Electrical Engineering and Computer Science, National United University, Miaoli, Taiwan from 2006 to 2009. Also he was a Chair professor in the Department of Computer Science and Information Engineering, National Taiwan University of Science and Technology, Taipei Taiwan from 2016 to 2023. Currently, he is a Chair professor in the Department of Computer Science and Information Engineering, Asia University, Taichung Taiwan. He has published more than 300 research papers and received many awards; especially, the Distinguished Research Award between 2004 and 2006 from the National Science Council in Taiwan; Outstanding IT Elite Award, in 2005; Outstanding EE Prof. Award, the Chinese Institute of Electrical Engineering; and the Outstanding Research and Invention Award between 2006 and 2008 from National Taiwan University of Science and Technology. He was also promoted to the Chair professor in National United University in 2008. His research interests include Information Security, Deep Learning, Biometric Recognitions, and Parallel Processing.