Analytics for Learning with Machines

• Applicant: Professorin Dr. Oleksandra Poquet
• Subject Area: General and Domain-Specific Teaching and Learning ; Human Factors, Ergonomics, Human-Machine Systems ; Education Systems and Educational Institutions ; Methods in Artificial Intelligence and Machine Learning
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 568938703

ALMA (Analytics for Learning with Machines) will create a foundation for understanding, supporting, and evaluating collaboration quality between learners and Large Language Models (LLMs) in higher education. As generative Artificial Intelligence permeates learning practices, future professionals need the skills to effectively collaborate with LLMs, beyond simple task offloading. To enable collaborative learning with LLMs, several challenges need to be addressed. First, the definition of collaboration quality needs to reflect to learner-LLM contexts surpassing existing pedagogies designed for human-human interaction. Second, LLMs need to be modified as their outputs do not exhibit collaborative behavior nor offer domain-specific reasoning. Third, methodologies for studying learner-LLM interactions need to move beyond a narrow focus on tool usage to enable insight about system-level elaboration processes. To overcome these challenges, ALMA will define collaboration quality for learner-LLM systems and develop a comprehensive framework that integrates pedagogical, technological, and analytical dimensions to support it. First, ALMA will re-articulate pedagogical foundations for collaborative learning with LLMs by engaging teachers in participatory co-design. Second, ALMA will develop a cutting-edge LLM infrastructure that affords collaborative behaviour by utilising hybrid fine-tuning techniques. Third, ALMA will innovate process-based analytics to evaluate collaboration quality by modelling how learner-LLM systems co-elaborate over time. The theoretical and practical contributions of the project will redefine how collaboration between learners and LLMs is understood, designed, and measured, while advancing research in Computer-Supported Collaborative Learning, Technology-Enhanced Learning, and Learning Analytics.

• DFG Programme: Research Grants
• International Connection: France
• Cooperation Partners: Emiliano Lorini, Ph.D.; Mar Pérez-Sanagustín, Ph.D.