Speakers

Keynote Speakers of 2024

 

Prof. Jinwen Ma


Peking University, China

 

   

Speech Title: The Improved Large Language Model for Mathematical Computing and Reasoning

Abstract: After the release of chatGPT and GPT-4, the transformer-based large language models have shown tremendous performance improvements and achieved great success in natural language processing and artificial intelligence. However, the key bottleneck problems of the large language model on its further developments and applications are the low ability of mathematical computing and reasoning. In this speech, we investigate how to improve the large language model to get the high performance of mathematical computing and reasoning by adding the continual learning mechanisms and inserting a system of inference rules. We can also improve the mathematical performance of the large language model by adopting and fusing a knowledge database in some way. As long as the large language model has the real ability of mathematical computing and reasoning, it can become powerful and applicable in the real scenarios and make the great-leap-forward development in artificial intelligence.

Biography: Jinwen Ma received the Ph.D. degree in probability theory and statistics from Nankai University, Tianjin, China, in 1992. Then, he joined the Department or Institute of Mathematics of Shantou University, Guangdong Province, China, and become a full professor in 1999. Since September 2001, he has joined the Department of Information Science at the School of Mathematical Sciences, Peking University, where he has served as the chair and full professor as well as a Ph. D. tutor in applied mathematics and now he is a full professor and PhD tutor in the Department of Information and Computational Sciences at the School of Mathematical Sciences, Peking University. During 1995 and 2004, he visited several times to the Department of Computer Science & Engineering, the Chinese University of Hong Kong as a Research Associate or Fellow. From September 2005 to August 2006, he was a Research Scientist with the Amari Research Unit, RIKEN Brain Science Institute, Japan. From September 2011 to February 2012, he visited as a Scientist to the Department of Systems Medicine and Bioengineering, The Methodist Hospital Research Institute, Houston, USA.
His main research interests include neural computation, machine learning, independent component analysis (ICA), computer vision, big data and large language models. He is the author or coauthor of more than 200 academic papers among which more than 60 papers were indexed by the Science Citation Index (SCI)－Expended. In fact, these papers have been cited over 4000 times. He has served as the Principal or Major Investigator for eleven national and three provincial or ministerial and two other scientific research grants as well as over 10 cross-sectional research projects. At present, he is the vice-director member of the Signal Processing Society in the Chinese Institute of Electronics (CIE) and a member on the editorial board of “Signal Processing (in Chinese)”. Moreover, he is the director of the Education Informationization Special Committee of China Chapter of International Information Study Society. He has served as a program committee member of several major international conferences such as ISNN, ICIC, ICONIP, ICSP. He was a co-chair of the program committee of 1999 Chinese Conference on Neural Networks and Signal Processing and the chair of the organization committee of the Third International Conference of Intelligence Science (ICIS 2018). He was selected in the 2017 AI Impact Scholars released by Ascemap and scholar.chinaso.com and the World’s Top 2% Scientists 2020 (Career Scientific Impact) released by Stanford University.

 

Assoc. Prof. Hoshang Kolivand


Liverpool John Moores University, UK

 

   

Speech Title: How Machine Learning is Reshaping Mixed Reality

Abstract: In this talk, we delve into the profound impact of ML on Mixed Reality, uncovering the latest advancements and groundbreaking innovations that are reshaping our digital experiences. From sophisticated real-time simulations to personalized virtual environments, explore how AI's integration with Mixed Reality is driving unprecedented immersion and transforming the way we perceive and interact with the virtual world. Join us as we unravel the limitless possibilities and implications of this transformative fusion.

Biography: Hoshang Kolivand is an Assoc. Prof in AI and Mixed Reality at Liverpool John Moores University (LJMU). With an MS degree in Applied Mathematics and Computer Science and a PhD and a Postdoc in Augmented Reality, he is a leading expert in these fields. As the Head of the Applied Computing Research Group at LJMU, Dr. Kolivand leads a team of over 35 researchers, focusing on AI and Augmented Reality. He has published extensively with over 170 papers in international journals and has presented at numerous conferences. Dr. Kolivand is a Senior Member of the IEEE and has served as a keynote speaker at more than 55 international conferences. He has organized over 30 conferences in AR, VR, AI, and HCI. In addition to his academic contributions, Dr. Kolivand has authored book chapters and several products which received over 14 awards for his work in Virtual Reality and Augmented Reality. As a dedicated researcher and educator, Dr. Hoshang Kolivand plays a significant role in advancing AI and Mixed Reality technologies, making valuable contributions to the field through his expertise and leadership.

 

Dr. Aminu Bello Usman
Head of the School of Computer Science

University of Sunderland, UK

 

   

Speech Title: Securing Tomorrow: Enhancing Biometric Image Privacy and Security through IoT and LPWAN Innovations

Abstract: "Securing Tomorrow: Enhancing Biometric Image Privacy and Security through IoT and LPWAN Innovations" the presentation explores advancements in securing biometric data within Internet of Things (IoT) ecosystems using Low Power Wide Area Networks (LPWAN). It focuses on addressing privacy and security challenges in biometric image transmission, storage, and processing. The presentation delves into how IoT and LPWAN technologies can be leveraged to create robust, scalable solutions for safeguarding sensitive biometric data, emphasising the need for innovative encryption techniques, secure protocols, and privacy-preserving mechanisms to protect against cyber threats and data breaches in an increasingly connected world.

Biography: Dr. Aminu Bello Usman served as the head of the School of Computer Science at the University of Sunderland. His research focuses on the Internet of Things (IoT), biometric security, applied AI, data privacy, trust, and user privacy. Dr. Usman is particularly passionate about IoT communication protocols, and his most recent works are on developing models, and frameworks that enhance user privacy and trust, addressing real-world security challenges of IoT and Edge computing.

 

Invited Speaker of 2024

 

Prof. Kwang Sik Chung


Korea National Open University, South Korea

 

   

Speech Title: Learning Contents Difficulty Analysis by Text Mining and Deep Learning Model on Learning Contents

Abstract: Despite the development of the online learning environment, it is difficult to estimate the learning level of the learner in the online environment, and various analysis methods for the learner are being studied in situations where direct interaction between the learning content and the learner occurs. However, the analysis of the difficulty level of learning content is a relative factor to the learners who encounter the learning content, and it needs to be analyzed in line with the learner's learning level. In addition, it has become necessary to determine the difficulty level of text-based learning content (textbooks, lecture notes, final exams, etc.) based on the learner's level.
In this study, we analyze the difficulty of text-based learning content and develop a system that links it with relatively relevant learning support services according to the learner's learning ability. For this purpose, learning-connected keywords are extracted by analyzing the prerequisite and follow-up subjects of a specific subject in the computer science curriculum. Focusing on the extracted learning connection keywords, the learning proximity for the preceding subjects is extracted, and the learning proximity for the subsequent subjects to be learned in the future is extracted. By analyzing the text of a specific subject, a learning-related ontology related to the keywords of the preceding subject is constructed, and based on this, a learning difficulty score is assigned to the learning contents related to the specific subject. Learners' learning responses to learning content with a high learning content difficulty score (keywords with high proximity to previous subjects) and learning content with a low learning content difficulty score are extracted together. Finally, the correlation between learning content difficulty and learning response is extracted to verify the effectiveness of the learning content difficulty analysis model. In addition, by extracting the difficulty level of text-based learning content, it is possible to classify learning content tailored to each learner, and through this, individualized learning support for each learner is possible. The results of the difficulty analysis of learning content can be used as basic data for test preparation, learner learning counseling system, and student competency strengthening system. In many other fields, it can be used as base data for various services through analysis of the difficulty of learning content.

Biography: Kwang Sik Chung received the Bachelor of Science (1992), Master (1995), and the Doctorate degrees (2000) in Computer Science and Engineering from Korea University, Seoul, Korea. Upon completing his degree, he worked as the research fellow at the Department of Computer Science at the University College London (UCL), London, United Kingdom from September 2002 to November 2003. Ever since returning back to Korea in 2005, he has been lecturing at the Department of Computer science at Korea National Open University (KNOU) as tenure track assistant professor.
His research interests include distributed systems, fault tolerant systems, and grid computing systems. He has been conducting various researches in the fields related to learning analytics, virtual experiments/practice learning contents system development for e-learning, and learning cloud construction as an international cooperative research and a visiting researcher at various universities for about 20 years since the beginning of 2000. He researched advanced technology in the fields of learning analytics and learning cloud development with a number of researchers.