The doctoral dissertation in the field of Computer Science will be examined at the Faculty of Science, Forestry and Technology, Joensuu campus and online.
What is the topic of your doctoral research? Why is it important to study the topic?
My research designs a human-AI collaborative adaptive learning system to support self-regulated learning (SRL) in online STEM education. Self-paced online learning offers flexibility but suffers from high attrition rates as it demands strong SRL skills that many learners lack. Existing adaptive learning technologies can help but are limited by their reliance on binary correctness data and machine-centric control, stifling learner autonomy essential for SRL.
My research addresses this critical gap by designing a collaborative partnership between learners and AI in a confidence-based adaptive practicing system that integrates learners' self-reported confidence and objective performance. This partnership leads to more accurate knowledge tracing, preserves learner agency, and actively develops metacognitive skills.
The ultimate goal is to transform online learning from a passive experience into an empowering journey that improves retention by cultivating capable, self-directed learners.
What are the key findings or observations of your doctoral research?
My research demonstrates that integrating learner confidence into adaptive algorithms significantly improves system performance. The novel ZPD-KT knowledge tracing model reduced prediction errors by 33% and increased pedagogically effective question selection by 25% compared to standard models. Furthermore, the AI-learner shared control approach successfully preserved autonomy and fostered metacognitive skills.
This work is novel for bridging metacognitive theory (e.g., Zimmerman's SRL) with computational innovation, dynamically operationalizing Vygotsky's Zone of Proximal Development.
For scientists, it offers a new and transparent ZPD-KT framework. For the public, it presents a practical prototype showing how AI can be a collaborative partner in education, empowering learners rather than just instructing them, which is vital for engaging online learning.
How can the results of your doctoral research be utilised in practice?
The results can be utilized in practice through the Confidence-based Adaptive Practicing web-based system, which can be integrated into existing online STEM courses. For educators and institutions, it offers a tool to reduce dropout rates by providing automated and personalized practice that identifies struggling learners. This gives instructors actionable insights to target interventions efficiently.
For learners, it provides an empowering learning tool that builds metacognitive skills and enhances learning awareness through confidence ratings and progress dashboards, enabling them to become more independent and self-regulated learners. The system’s design (particularly its transparency and shared control) also serves as a best-practice model for EdTech researchers to develop ethical and human-centric AI educational tools.
What are the key research methods and materials used in your doctoral research?
My doctoral research employed a Design Science Research (DSR) methodology, following an iterative, five-phase process to create and evaluate a practical solution.
The key phases were:
1. Problem Explication: Analyzing literature and institutional data, and conducting exploration interviews to identify key barriers in online STEM learning.
2. Requirements Definition: Deriving functional needs of the adaptive learning system from the problems.
3. Artifact Design & Development: Designing the six-module CAP prototype and its core innovation, the ZPD-KT knowledge tracing framework.
4. Demonstration: Building a functional web application to showcase the system's adaptive logic.
5. Evaluation: Conducting a large-scale simulation with 8,000 synthetic learners to rigorously benchmark the ZPD-KT model's performance against traditional standards like Bayesian Knowledge Tracing (BKT). This method allowed for controlled, ethical, and precise validation of the artifact's efficacy.
The doctoral dissertation of Hongxin Yan, MEng, entitled Human-AI integrated adaptive practicing to foster self-regulated learning in online STEM education will be examined at the Faculty of Science, Forestry and Technology, Joensuu campus. The opponent will be Professor Peter Reimann, The University of Sydney, and the custos will be Professor Markku Tukiainen, University of Eastern Finland. Language of the public defence is English.
For more information, please contact:
Hongxin Yan, [email protected]
- Public examination
- Dissertation (PDF)
- Photo (coming)
