16th International Cognitive Load Theory Conference 2024

'Intelligence' is a commonly used term that is never properly defined. We assume humans are the most intelligent animals with other animals having lesser amounts. The advent of artificial intelligence has given rise to suggestions that we may have developed something that will eventually exceed us in intelligence. The cognitive architecture used by cognitive load theory can provide us with a base for considering intelligence. That base suggests firstly, that rather than being fixed, intelligence is heavily dependent on the variable contents of long-term memory and secondly, in evolution by natural selection, we already are faced with an intelligent system that vastly exceeds our own intelligence.

John Sweller, Emeritus Professor of Educational Psychology, University of New South Wales

John's research is associated with cognitive load theory. The theory is a contributor to both research and debate on issues associated with human cognition, its links to evolution by natural selection, and the instructional design consequences that follow. Based on hundreds of randomised, controlled studies carried out by many investigators from around the globe, the theory has generated a large range of novel instructional designs from our knowledge of human cognitive architecture. Based on any commonly used citation index, the work has been cited on more than 25,000 occasions.

How do educators harness the best of explicit instruction and discovery learning? Load reduction instruction (LRI) is an instructional strategy aimed at addressing this question. Drawing on key concepts and evidence from cognitive load theory and information processing models, LRI seeks to ease the cognitive burden on students so they can learn effectively. Initially, LRI involves explicit instruction. Then, as students develop fluency and automaticity in knowledge and skill, LRI moves onto less structured approaches such as guided discovery-, problem-, and inquiry-based learning. In this presentation, Andrew will explain LRI, share recent research findings, and identify some of the methodological approaches and opportunities LRI affords researchers in education and psychology.

Andrew J. Martin, Scientia Professor, Professor of Educational Psychology, and Chair of the Educational Psychology Research Group, School of Education, University of New South Wales

In addition to his substantive position with The University of New South Wales, Andrew is an Honorary Research Fellow in the Department of Education at the University of Oxford and a registered psychologist with the Psychology Board of Australia. He specialises in student motivation, engagement, learning, and quantitative research methods. He is a consulting editor for Psychological Review, Journal of Educational Psychology, and Educational Psychology and serves on numerous international editorial boards including Educational Psychologist, Learning and Instruction, Contemporary Educational Psychology, British Journal of Educational Psychology, Learning and Individual Differences, and Journal of Experimental Education. He is a Fellow of the American Psychological Association and the American Educational Research Association.

I approach this topic from the perspective of the role of models in instructional design. ChatGPT, for instance, is a Large Language Model (LLM), so questions such as these arise: What is an LLM a model of? How is the modelling done? How does the model create value for instructional designers, learners, and researchers? I discuss in what sense LLMs model the knowledge to be learned - the 'content'- because that seems to be their primary function in supporting instructional design. I also discuss how LLMs might serve as models of learners and learning, called the "student model" in intelligent tutoring systems. I critically compare the similarities and differences to more established content and student modelling methods, namely Semantic Models, Bayesian Network Models, and Cognitive Models. By the end of the presentation, it should become clearer to what extent LLMs contribute to insights and create value beyond other approaches to educational modelling.

Peter Reimann, Professor of Education, Co-Director, Centre for Research on Learning and Innovation, Sydney School of Education and Social Work, The University of Sydney

Peter has been a Professor of Education at The University of Sydney since 2003. He has a PhD in psychology from the University of Freiburg, Germany. Having initially trained as a cognitive psychologist with a general interest in technology-augmented and technology-mediated learning, Peter has frequently applied computational modelling methods to advance learning research. His main research area is Computer-supported Collaborative Learning. One of his contributions to CSCL was introducing process-mining methods for analysing trace data. Also relevant to CSCL research is his work (with Jacobson and Kapur) on the role of complexity theory for theories of human learning. Currently, he and his PhD students are conducting research employing semantic technologies such as Knowledge Graphs, to support peer tutoring and learning from argumentation. During his career, Peter has directed two research centres: The Centre for Research on Computer-supported Learning & Cognition (CoCo) with Peter Goodyear, and the Centre for Research on Learning and Innovation (CRLI) with Lina Markauskaite.