Prof. Brendan Flanagan, Ritsumeikan University, Japan

Bio: Brendan Flanagan is a Professor at the College of Information Science and Engineering and the Graduate School of Information Science and Engineering, Ritsumeikan University. His research focuses on AI in education, learning analytics, and educational data science, with recent work exploring LLM-supported language learning, explainable educational recommender systems, and the application of blockchain in education. A long-standing advocate for open educational data, he has hosted educational data challenges at leading international conferences for more than eight years and recently co-authored a large-scale survey of open datasets in learning analytics, published in ACM Transactions on Knowledge Discovery from Data. He is currently Principal Investigator of several government-funded projects investigating knowledge and learning process analysis and explainable predictions in learning systems and was a member of a large-scale NEDO-funded research project on educational symbiotic AI systems.

 

Talk Title: Educational Data Science: Changes and Challenges in Learning

Abstract: As education technology enters the age of large-scale digital platforms, the field faces informatics challenges that resonate with broader data science and AI research: how to capture heterogeneous traces of human activity, design interoperable infrastructures, and develop transparent models that scale across domains. In this keynote, I argue for a research agenda that positions educational data as a frontier testbed for advancing informatics: demanding innovations in interoperability, explainability, and real-world deployment. I will present recent work on learning analytics infrastructure and standards that captures learning traces and enables cross-institutional data sharing while respecting privacy; explainable recommendation systems that allow learners to interrogate model outputs, together with human-in-the-loop approaches to knowledge modeling that balance algorithmic performance with expert oversight; and real-world deployment in which learners analyze their own learning logs as part of data science education in Japan. I will then turn to shared data as the foundation of reproducible educational data science, drawing on a recent systematic survey of open datasets in learning analytics, covering 1,125 papers and 172 datasets across the leading conferences in the field, that reveals how rarely data is shared and proposes actionable PRACTICE guidelines, and on synthetic data generation that enabled an international educational data analysis challenge without exposing private learner data. Drawing from experience organizing international data challenges and embedding systems in production environments, I will discuss lessons learned in domain adaptation, evaluation beyond benchmark metrics, and balancing algorithmic performance with human interpretability. These challenges parallel those faced across other areas of informatics: ensuring that models are not only accurate, but also trustworthy, explainable, and adaptable to dynamic, real-world contexts.

 

Prof. Jon Dron, Athabasca University, Canada

Bio: Jon Dron is a full professor and Associate Dean, Learning & Assessment for the Faculty of Science and Technology at Athabasca University, Canada. He researches in the fields of education, technology and online learning, to which he has been making theoretical, practical, and technical contributions since the 1990s. He is the author of three books on learning and technology: How Education Works: Teaching, Technology & Technique (2023); Teaching Crowds: Learning & Social Media (2014); and Control & Constraint in E-Learning: Choosing When to Choose (2007). He has been a UK National Teaching Fellow since 2005. Prior to becoming an academic, Jon had careers in IT and as a singer of swing. He has a BA in philosophy, an MSc in information systems, a postgraduate diploma in education, and a PhD in learning technology. He teaches and researches at a distance from his home in Vancouver. His blog and personal website is at https://jondron.ca.

 

Talk Title: Why educational transformation is hard, and how to make it easier

Abstract: Changing parts of our higher educational systems is easy, but truly transforming them is very challenging, thanks to their many roles both internally and in the societies in which they are situated, the many stakeholders with competing needs, and the deep path dependencies brought about by their long histories, all of which conspire to resist substantive change. In this talk, drawing theory from my book “How Education Works: Teaching, Technology, and Technique" I will consider the problem from a technological perspective, using a very broad definition of the term “technology” which embraces poetry and pedagogical methods as much as pottery and potentiometers. I will use this to explain why higher education has, till now, been so resistant to transformation and how, thanks to widespread availability of generative AI, that transformation is bound to occur whether we like it or not. I will describe ways that all stakeholders, from students to education ministers and university presidents, can turn this into a positive transformation by properly understanding the unique nature of generative AIs as fellow tool-users but not fellow beings-in-themselves. Along the way you will learn why what we usually think of as the best ways to teach are actually only necessary because of the problems caused by the ways we choose to teach and assess, why generative AI is such a threat unless we do transform, how it can help us to break out of that vicious cycle, and how this changes the roles of learners and those who support their learning everywhere.

 

Prof. Chengjiu Yin, Kyushu University, Japan

Bio: Dr. Chengjiu Yin is currently a Vice Director of the Research Institute for Information Technology at Kyushu University, Japan. He received the Early Career Research Award from APSCE in 2019. Prof. Yin serves as an Associate Editor of Interactive Learning Environments (SSCI). He has also served as a Guest Editor for the Journal of Distance Education Technologies (EI) on three occasions and for Interactive Learning Environments (SSCI) on four occasions. He has been invited to deliver over 20 speeches at various academic conferences and events. He is a member of the Editorial Board for the Journal of Educational Technology & Society (ET&S) and for Computers and Education: Artificial Intelligence. Prof. Yin emphasizes the continuous adoption and development of new teaching methods, materials, and technologies grounded in scientific insights. His research focuses on Educational Data Mining, Learning Analytics, and AI in Education. Currently, his primary goal is to explore how AI can be used to design learning environments that enhance knowledge sharing, awareness, and creation. He has proposed the concept of "Learning Success" for the era of AI in education and leads research on the transformation of educational goals.

 

Talk Title: Learning Behavior Analysis Based on Digital Book Learning Logs

Abstract: Digital books have been increasingly adopted in educational settings, providing students with flexible access to learning materials while enabling the collection of detailed learning logs. These logs provide valuable information about students’ interactions with learning materials and offer opportunities to better understand their learning processes. This topic aims to analyze student learning behavior using learning logs collected from a digital book system. The log data include various learning activities, such as page navigation, reading duration, highlighting, bookmarking, and note-taking. By analyzing the frequency, timing, and sequential patterns of these activities, this study investigates how students interact with digital learning materials and identifies different patterns of learning behavior.