Description

In today’s dynamic educational environment, learners transition to senior school or higher
institutions of learning face complex decisions about selecting pathways that align with
their competencies, interests, and future career goals. Despite the increasing use of AI in
educational recommender systems as guidance tools, most existing models are often
generic, non-adaptive, and lack transparency. This research is motivated by the need to
empower learners with a system that adapts and evolves with their behavior over time,
personalizes pathway recommendations based on long-term progression and feedback,
and explains the recommendations in a way that learners and educators can understand
and trust. By combining RL and XAI, the system aims to balance personalization and
adaptability with interpretability, ensuring trust and actionable feedback for both learners
and educators. There is therefore a need for a hybrid approach that leverages the
exploratory learning power of reinforcement learning (for adaptability and personalization)
and the transparency of explainable AI to deliver a truly learner-centric pathway guidance
tool.

Main Objective

To develop and evaluate a personalized, adaptive, and interpretable Pathway
recommendation system for transitioning students by integrating reinforcement learning
and explainable AI techniques.

Specific Objectives:

1. To design a reinforcement learning-based recommendation framework that
   dynamically learns and adapts to individual learner profiles, preferences,
   competences, and performance patterns.
2. To integrate explainable AI techniques into the recommendation model to generate
   human-understandable explanations for suggested learning pathways.
3. To develop a prototype of the hybrid recommendation system with an interactive
   interface that supports exploration and understanding of recommendations by
   learners and educators.
4. To evaluate the effectiveness of the developed system