The K-CUBE PET platform is an innovative educational technology project designed to enhance teaching and learning through personalized and interactive experiences. Its key innovation lies in its ability to create adaptive learning pathways by analyzing student interactions and performance data using machine learning and natural language processing. This system dynamically updates content based on student progress, ensuring a tailored learning experience.

The design concept revolves around real-time data collection and analysis, enabling the system to adapt to individual student needs effectively. Technical principles include advanced generative AI algorithms that process behavioral patterns and generate customized educational material, fostering engagement and understanding.
By addressing current limitations in education, such as one-size-fits-all teaching methods, K-CUBE PET aims to unlock new possibilities for both students and educators. Its potential impact is significant, with the possibility of improving student outcomes through personalized learning and increasing educator efficiency by providing insights into the progress of each individual student.

The K-CUBE PET versatility allows it to be applied across various educational settings, making it a promising tool for modernizing the higher education. By leveraging cutting-edge technologies, K-CUBE PET presents a forward-thinking approach that could revolutionize how we teach and learn.

The K-CUBE PET is an immersive educational system that uses machine learning and generative AI in a CAVE VR/AR environment. The goal is to create personalized learning experiences for students by adapting content based on their interactions and performance.

The inspiration came from the need to make education more engaging and effective. Traditional methods might not cater to individual learning styles, so using technology could bridge that gap. We noticed that students learn better when material is tailored to them, hence the idea of adaptive pathways.

The immersive aspect with the CAVE VR/AR adds a layer of engagement that textbooks or online courses cannot match. It makes learning more interactive and memorable. We have measured the factors that contribute to the learning outcomes using K-CUBE PET. The results confirm this hypothesis. Furthermore, integrating AI avatars that can assist and guide students provide real-time help.

We believe that combining personalized adaptive systems with immersive tech would lead to better student outcomes. By visualizing knowledge graphs in 3D, students could understand complex concepts more easily.

In tertiary education, there is a need for more interactive tools that go beyond lectures and readings. The success of the K-CUBE PET project relies on how well it engages students and make learning more effective through immersion, presence, engagement, and trust.

The K-CUBE PET project leverages cutting-edge LLMs and Knowledge Graphs to revolutionize content creation in learning media. By integrating LLMs with KGs, the system can dynamically retrieve accurate, up-to-date information from structured knowledge bases and generate context-rich, engaging text for news articles, reports, discussions, meetings, videos, and marketing materials. This integration ensures that content is not only well-written but also enriched with reliable data.

High-quality knowledge graphs that are continuously updated with the latest facts and figures require continuous management and maintenance. Additionally, computational resources such as powerful servers, supported by the latest GPUs and TPUs within a cloud infrastructure will be necessary to run the LLMs efficiently. We will also need datasets on user preferences and behaviors to personalize content delivery effectively. In our current prototype we leveraged our department teaching curriculum and computing resources to implement the first version of the K-CUBE PET system.

The evaluation focuses on metrics such as student satisfaction and performance, demonstrating its effectiveness in real-world settings. Challenges include maintaining up-to-date knowledge graphs and ensuring scalability for diverse user needs. Usability is prioritized with user-friendly interfaces for both students and educators. Ethical considerations ensure equitable access and prevent bias, making the K-CUBE PET platform a comprehensive tool for enhancing learning outcomes inclusively.

To validate market demand, we plan to conduct pilot studies with media companies and content creators. These pilot studies will assess the system's ability to generate high-quality content compared to traditional methods. We have already conducted extensive learning experiments with our students and have documented the results in ten published academic papers. We will further gather feedback from users and analyze metrics such as time-to-content creation, cost efficiency, and audience engagement. Surveys and focus groups with end-users will also provide insights into the effectiveness of personalized content in meeting their needs.

NA

The K-CUBE PET platform is an innovative educational technology project that leverages advanced technologies such as multi-agent LLMs to create personalized and interactive learning experiences. The system's core functions include real-time data collection, analysis, and dynamic content generation, which are supported by our generative AI teaching assistant, or tutor.

The integration of LLM allows processing vast amounts of student interaction data, identifying areas where learners may struggle or excel. This analysis informs the generation of customized content, fostering engagement and understanding. LLM agents enhance this by enabling natural language interactions, making learning more intuitive and responsive to student queries or feedback.

The K-CUBE PET system represents an innovative and creative solution to the problem of one-size-fits-all education by dynamically adapting learning experiences through advanced technology. Our platform stands out in the following aspects:

1. Dynamic Adaptation: it utilizes machine learning and natural language processing to analyze student interactions in real-time, adjusting content based on individual progress.
2. Personalization: it offers tailored educational material that caters to each student's needs, enhancing engagement and efficiency by addressing their specific learning pace and style.
3. Insightful Analysis: it goes beyond right/wrong answers by examining behavioral patterns like time spent and question skips, providing educators with detailed insights for targeted interventions.
4. Versatility: it is applicable across various educational settings, from K-12 to corporate training, broadening its impact and adoption potential.
5. Interactive Learning: it incorporates conversational interfaces, allowing students to interact naturally, making learning feel more personalized and less static.
6. Continuous Improvement: it evolves with each interaction, refining adaptive pathways for enhanced effectiveness over time.

In essence, the K-CUBE PET revolutionizes education by transforming it into a dynamic, data-driven journey through an immersive 3D knowledge graph that continuously adapts to individual needs, promising improved outcomes for both students and educators.

Scalability requires leveraging cloud infrastructure, scalable databases, caching, load balancing, efficient content generation, user experience monitoring, and future-proofing through modularity. Each of these strategies will help the K-CUBE PET scale effectively while addressing potential bottlenecks. To ensure the K-CUBE PET system is scalable and addresses potential bottlenecks, the following strategies are employed:
1. Cloud-Based Infrastructure: Utilizes scalable cloud services to dynamically adjust resources based on demand, ensuring efficient handling of a growing user base.
2. Scalable Databases: Employs NoSQL databases to manage large volumes of interaction data efficiently, accommodating growth without performance degradation.
3. Caching Mechanisms: Implements caching to reduce server load and accelerate data retrieval, enhancing overall system responsiveness.
4. Load Balancing: Distributes traffic across multiple servers to prevent overload, ensuring consistent performance during peak usage.
5. Optimized Machine Learning Models: Develops lightweight models to maintain speed and accuracy, crucial for real-time adaptive learning.
6. Efficient Content Generation: Uses optimized algorithms and parallel processing to quickly generate tailored content as the user base expands.
7. User Experience Monitoring: Conducts regular testing with real users to identify and address performance issues, maintaining an intuitive interface.
8. Modular Architecture: Designs a system that allows for easy addition of new features without disrupting existing functionalities, ensuring future growth.
9. Regular Updates and Maintenance: Implements ongoing improvements to maintain system efficiency and reliability, anticipating future scalability needs.
These strategies collectively ensure the K-CUBE PET system remains scalable, efficient, and user-friendly as it grows

The K-CUBE PET platform addresses environmental sustainability by minimizing reliance on physical resources through its digital platform, thereby reducing paper waste and lowering the carbon footprint associated with traditional educational materials. It fosters long-term user engagement by employing adaptive learning pathways that adjust content dynamically based on student interactions and performance, incorporating interactive elements like real-time feedback to maintain interest.

The K-CUBE PET system's ability to analyze data in real-time means it can identify areas where a student is struggling early on. This proactive approach can help educators intervene, preventing students from falling too far behind. Early intervention is crucial for keeping students engaged and on track, which contributes to better educational outcomes overall.

Engagement is another key factor. When students are interested and challenged appropriately, they are more likely to succeed. The K-CUBE PET interactive content keeps the students motivated, which can lead to improved academic performance and a more positive attitude towards learning.

Looking at broader social goals, equity and inclusion are central. By addressing the unique needs of each student, the system supports these goals by ensuring that all students have an equal opportunity to succeed, regardless of their background or initial skill level.

Inclusivity is also promoted because the system can accommodate different learning styles and languages if designed that way. This makes education more accessible to a wider range of students who might otherwise feel excluded.

Our detailed experiments have been published in reputable academic journals and conferences. These show that metrics such as cognitive presence, engagement and trust contribute significantly to learning outcomes and cognitive retention. When students are engaged with content tailored to their needs, they learn better and stay motivated. Success would be measured by improvements in test scores, student feedback, and perhaps increased retention rates. Our immersive approach makes learning more memorable and interesting. Adaptive generative AI systems can improve educational outcomes by catering to each student's needs, potentially leading to better test scores, higher satisfaction, and increased retention rates. By its own nature this will evolve with the needs of the higher education community and ensure responsiveness which leads to higher social impact in a rapidly changing economic environment.

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.