Open Category Submission

343 – AI-Enhanced Adaptive Learning Platform for Multiple-Choice Assessments

AI-Enhanced Adaptive Learning Platform for Multiple-Choice Assessments is an innovative online platform designed to enhance the learning experience through personalized and adaptive multiple-choice question (MCQ) assessments powered by generative AI. The platform allows teachers to create and manage a dynamic question bank that students can access to practice MCQs across various topics.

Key Innovations:
1. Teacher-Prepared Content: Teachers begin by inputting a set of high-quality MCQs into the platform. This ensures that the foundational content is aligned with curriculum standards and tailored to the specific learning objectives of different subjects.
2. AI-Enhanced Question Generation: Utilizing AI technology, teachers can generate additional questions based on existing ones. This feature enables educators to expand the question bank efficiently while maintaining quality and relevance. The generated questions are then submitted for teacher approval before being added to the question bank, ensuring that all content meets educational standards.
3. AI-Powered Explanations: Each question includes an AI-generated explanation that clarifies the correct answer, promoting deeper understanding and retention of concepts. This aspect supports students in grasping complex topics and encourages self-directed learning.

Technical Principles:
The platform employs machine learning models to analyze student responses and performance trends. The initial phase focuses on teacher-prepared questions and AI-generated content, while future iterations will explore more advanced adaptive learning features. A reporting system allows students and teachers to flag inaccuracies in questions or explanations, fostering a culture of continuous improvement and ethical AI use.

Potential Impact:
This platform aims to transform traditional assessment methods by making them more personalized, engaging, and effective. By leveraging AI, it enhances accessibility and inclusivity in education while providing educators with valuable insights into student performance through detailed analytics. This holistic approach fosters a deeper understanding of subject matter, promotes equity in learning, and prepares students for future academic challenges.

In today’s educational landscape, traditional assessments often fail to accommodate the diverse needs and learning styles of students. Many learners struggle with static, standardized question sets that do not adapt to their individual strengths and weaknesses, leading to disengagement and lower performance. Meanwhile, teachers face the time-consuming challenge of creating varied, high-quality assessment materials aligned with the curriculum.

To address these issues, I initially developed an online MCQ practice platform that allows students to select topics, practice questions, and track their performance. This tool helped streamline assessment delivery and provided valuable insights into student progress. However, while useful, it highlighted a persistent problem: the difficulty of continuously expanding the question bank to keep practice engaging, personalized, and relevant without excessive manual effort.

Inspired by this challenge, I decided to integrate AI capabilities into my existing platform. The core hypothesis is that combining teacher-prepared questions with AI-generated content and explanations—produced in manageable, teacher-controlled batches—will significantly enhance student engagement and learning outcomes. Teachers can seed high-quality questions, then periodically use generative AI to expand the question bank and create tailored explanations, all subject to their review and approval.

This approach combines human expertise with AI efficiency, enabling a dynamic, diverse, and curriculum-aligned assessment experience without the prohibitive costs of real-time AI deployment. It also empowers teachers to maintain quality control and ethical oversight.

By evolving my platform in this way, I aim to create a scalable, sustainable solution that addresses gaps in current assessment practices, fosters personalized learning, and supports diverse student needs—ultimately contributing to a more inclusive, motivating, and effective educational environment.

Our platform builds upon an existing web-based MCQ practice system developed using open-source technologies such as PHP, JavaScript, and MySQL. This foundation ensures stability and ease of maintenance while enabling the integration of new AI capabilities.

To generate additional questions and explanations, we will leverage locally hosted open-source large language models (LLMs) where feasible, running on a personal computer with a decent GPU. Instead of costly, real-time AI APIs, we will adopt a batch-processing approach: teachers periodically select seed questions and run AI generation offline or during low-demand times. This method minimizes hardware and financial requirements while still expanding the question bank efficiently.

Core functionalities include:
- Teacher-driven question input and curation
- Batch AI-assisted generation of new MCQs and explanations, reviewed by educators before deployment
- A student practice module with topic selection and performance tracking
- An analytics dashboard to monitor student progress and identify learning gaps
- A reporting system for teachers and students to flag problematic questions or explanations

To ensure a positive user experience, the platform offers a simple, intuitive interface for both teachers and students. Teachers maintain control over content quality, while students benefit from a continually refreshed, personalized question bank and clear explanations that support independent learning.

We will evaluate effectiveness through key metrics such as:
- Student engagement rates (usage frequency, session duration)
- Improvement in student scores over time
- Teacher approval rates of AI-generated content
- Rates of reported errors to guide iterative improvements
- User satisfaction surveys

The AI-Enhanced Adaptive Learning Platform for Multiple-Choice Assessments integrates existing MCQ capabilities with innovative AI features. The functional architecture consists of three main components: the Question Bank Management, Interactive Question & Answering, and the Enhanced Analytics Dashboard.

Functional Architecture and Technical Workflow:
1. Question Bank Management: Teachers input and curate a diverse question bank using a user-friendly web interface. This system utilizes a MySQL database to store both teacher-prepared and AI-generated questions.
2. AI Question Generation: Leveraging natural language processing (NLP), this module generates additional questions based on existing teacher-prepared questions. Teachers will review and approve generated questions before they are added to the MySQL database.
3. Enhanced Analytics Dashboard: The existing dashboard will be upgraded to include AI features that analyze student performance data, providing predictive insights to identify at-risk students and recommend personalized learning paths.

Implementation Process:
The implementation process is structured into two phases:
Phase 1: Completed. The foundational MCQ practice platform is established with teacher-prepared questions.
Phase 2 (Next 5 weeks): Develop the interface for AI-generated questions and incorporate machine learning models into the Analytics Dashboard.

Performance Metrics:
To assess the effectiveness of the AI system, the following metrics will be tracked and analyzed:
1. User Engagement Rates: Frequency of user logins and interactions with AI-generated questions, measured through daily active users (DAU), session duration, and completion rates.
2. Question Approval Rates: Percentage of AI-generated questions reviewed and approved by educators, reflecting alignment with curriculum standards and instructional quality.
3. Incorrect Question Reporting Rates: Proportion of AI-generated questions flagged as incorrect, ambiguous, or pedagogically unsound by teachers, indicating areas for model refinement.
4. Student Performance Improvement: Comparative analysis of student assessment scores before and after implementation of the AI tool, demonstrating measurable academic progress.

Our AI-Enhanced Adaptive Learning Platform tackles the rigidity of traditional assessments through a truly innovative, synergistic approach: teacher-AI collaboration for content generation. Instead of relying solely on static, manually-created questions or purely autonomous AI, our platform empowers educators to input foundational questions, which then serve as high-quality seeds for our generative AI. This AI module, leveraging NLP, creatively expands the question bank, offering diverse variations and targeting specific learning nuances.

The core innovation lies not just in using AI, but in the integrated human-in-the-loop workflow. Teachers are central; they review, refine, and approve every AI-generated question before it reaches students. This creative safeguard ensures pedagogical integrity, relevance, and quality, directly addressing concerns about AI-generated content reliability, and builds trust. This process transforms content creation from a laborious, static task into a dynamic, scalable, and collaborative one.

This hybrid model’s effectiveness is significantly enhanced by its ability to generate a vast, nuanced question bank, which is the bedrock for truly personalized and adaptive learning experiences. Furthermore, the planned inclusion of AI-generated explanations for each question offers immediate, tailored feedback, helping students grasp concepts more deeply. Our enhanced analytics, using AI to provide predictive insights, moves beyond simple scoring, offering educators actionable data to identify at-risk students and suggest personalized interventions. This blend of teacher-guided AI content generation, rigorous quality control, and intelligent analytics provides a creative, effective solution that directly combats student disengagement and supports diverse learning needs by making assessment a more dynamic, responsive, and supportive part of the learning journey.

Our platform adopts a practical, resource-efficient approach centered on batch-based AI content generation. Rather than generating new AI content in real-time for every student, the system enables teachers to periodically run AI models offline or during low-demand periods (e.g., once per term or per unit). This approach minimizes hardware requirements and operational costs, making it feasible to operate with a modest GPU-equipped personal computer instead of expensive servers or paid API services.

The platform’s architecture remains modular and cloud-compatible, allowing future migration to scalable cloud infrastructure when resources permit. Question bank management and analytics rely on lightweight, optimized code and database queries, ensuring smooth performance with limited hardware.

For sustainability, we favor efficient workflows that reduce computational load by:
- Reusing and refining previously generated AI content,
- Caching AI outputs for multiple uses,
- Encouraging human-in-the-loop validation to prioritize quality over quantity.
- Long-term engagement is supported through a teacher-centered workflow: educators curate and approve AI-generated questions in manageable batches, maintaining content quality and curriculum relevance. For students, this ensures a continually refreshed but reliable question bank that supports personalized practice without overwhelming system resources.

Our AI-Enhanced Adaptive Learning Platform is designed to address the social issue of unequal access to quality education, particularly for students from disadvantaged backgrounds. By providing personalized and adaptive learning experiences, our platform aims to bridge the gap in educational outcomes and enhance the lives of its primary beneficiaries - students and teachers.

Our solution aligns with broader social goals such as equity and inclusion by:
- Providing equal access to quality educational resources for students from diverse backgrounds
- Catering to different learning styles and abilities, ensuring that no student is left behind
- Empowering teachers to create inclusive and effective learning environments

To measure the social impact of our solution, we will track the following metrics:
- Increase in student engagement and motivation
- Improvement in academic performance, particularly for students from disadvantaged backgrounds
- Reduction in teacher workload and stress
- Increase in teacher satisfaction and confidence in using technology

To ensure responsiveness to the evolving needs of the community, we will:
- Regularly collect feedback from students, teachers through surveys and focus groups
- Continuously monitor and analyze user data to identify areas for improvement and inform future development

Personal Information Collection Statement (PICS):
1. The personal data collected in this form will be used for activity-organizing, record keeping and reporting only. The collected personal data will be purged within 6 years after the event.
2. Please note that it is obligatory to provide the personal data required.
3. Your personal data collected will be kept by the LTTC and will not be transferred to outside parties.
4. You have the right to request access to and correction of information held by us about you. If you wish to access or correct your personal data, please contact our staff at lttc@eduhk.hk.
5. The University’s Privacy Policy Statement can be access at https://www.eduhk.hk/en/privacy-policy.

• I have read and agree to the competition rules and privacy policy.