Moses Omopekunola Presents AI-Driven Psychometrics Research at IMPS 2026
As the largest annual gathering of researchers working at the intersection of statistics, machine learning, education, and psychology, IMPS provides a platform for sharing the latest research and discussing emerging trends in psychometrics. During the conference, Moses Omopekunola, Research Assistant at the Centre for Psychometrics and Measurement in Education, attended sessions on Automatic Item Generation (AIG) using Large Language Models (LLMs), Item Response Theory (IRT), Cognitive Diagnostic Models (CDMs), and Computerized Adaptive Testing.
Among the highlights of the programme were presentations on new statistical approaches to Cognitive Diagnostic Models by Professor Yuqi Gu (Columbia University), the use of Cognitive Diagnostic Models to evaluate large language models, and agentic AI frameworks for automated item generation presented by Rudolf Debelak (University of Zurich/EPFL) and Rasmus Jensen. Together, these sessions reflected the rapidly growing role of artificial intelligence in modern psychometrics.
Listening to these different presentations, two things stood out for me. The first is that research cannot be confined to one place: I saw that ideas our research group at HSE University had been discussing and planning to implement were already being explored elsewhere. At the same time, I was pleased to see that our research is up to international standards. The second is that AI is penetrating every fabric of research, and psychometrics is not left behind. I saw numerous presentations on the application of AI across many aspects of our discipline,” Moses shared his impressions.
During the session on AI, Validity, and Reliability, Moses presented the joint study "Agentic Large Language Models for Multimodal Physics Item Generation and Automated Item Review." He also introduced an R package developed by the Centre for Psychometrics and Measurement in Education, which attracted considerable interest from conference participants. The presentation prompted further discussions after the session, with researchers asking about the software and possibilities for future collaboration. As Moses believes, this will lead to future collaborations between HSE University and other leading universities around the world.
I want to thank the Faculty of Social Sciences and the Institute of Education, HSE University, for making my participation possible. My special gratitude goes to Ksenia Tarasova, the Director of our Centre, the Centre for Psychometrics and Measurement in Education for sponsoring my attendance at this great conference. I am also deeply grateful to my academic advisor, Elena Yu. Kardanova, for her support.”
*The next International Meeting of the Psychometric Society (IMPS) is scheduled for July 2027 at the University of British Columbia in Vancouver, Canada.
