FST Graduate Research Topic

Research Topic: Student Models that Support Intelligent Educational Systems
Suggested By: Oscar Lin
Last Update: May 14th, 2025

A "student model" in intelligent educational systems (IES) is a dynamic, computer-based representation of a learner’s current knowledge, skills, and other relevant characteristics. It serves as the system’s internal profile of the student, constructed and updated based on the learner’s interactions, behaviors, and performance data within the educational environment. The student model not only tracks what a student knows or struggles with, but can also include affective and motivational states, learning preferences, and background information. The main purpose of a student model is to enable personalized instruction: by maintaining an up-to-date record of each learner’s strengths, weaknesses, and learning progress, the system can adapt content, feedback, and instructional strategies to better meet individual needs. For example, if the model detects a misconception or lack of mastery in a particular topic, it can trigger targeted remediation or suggest additional practice. Current research in student modelling addresses a variety of challenges and innovations. For your MSc IS essay/project/thesis or BSc CIS final year project, you can choose one of the research topics on Student Modelling in Intelligent Educational Systems - Knowledge Tracing and Skill Mastery:    Techniques for tracking a student’s mastery of specific knowledge components over time, such as Bayesian Knowledge Tracing and deep learning approaches.   - Affective and Motivational Modelling:  Incorporating emotional states (e.g., engagement, frustration) and motivation into the student model to improve adaptivity and support.   - Automatic and Real-Time Modelling:   Using machine learning to automatically infer student characteristics (such as learning style, engagement, and emotion) from behavioral and physiological data, enabling real-time adaptation.   - Open Learner Models:   Making the student model transparent and accessible to learners themselves to promote self-regulation and reflection.   - Personalization and Adaptive Interventions:    Designing models that inform the system when and how to intervene, such as providing hints, feedback, or adjusting the difficulty of learning materials.   Implementation:     - Tools: Python (for graph construction, AI algorithms) and MongoDB (for scalable data storage).  - Support: Financial assistance may be available for students contributing to this project.     By grounding the design in Prof. Lin’s formative assessment framework and advancing it with knowledge graph technology and new AI tools, this work seeks to create more dynamic, interpretable, and effective intelligent tutoring systems.