Course Universe

Inspiration

Students often learn from disconnected textbooks, lecture videos, and presentation slides. Understanding how concepts connect, identifying missing prerequisites, and deciding what to study next can be challenging.

We envisioned Course Universe as a new way to explore knowledge: instead of treating learning materials as separate documents, we transform them into an interactive universe where every concept connects to related ideas and its original source.

What it does

Course Universe is an AI-powered personalized learning companion designed to turn course materials into an interactive knowledge map.

• Interactive Knowledge Galaxy: Visualizes topics, subtopics, and relationships between concepts.
• Source-Connected Learning: Connects explanations to their supporting textbook pages, slide numbers, or lecture timestamps.
• Knowledge Gap Detection: Identifies potentially weak concepts and missing prerequisites using assessment results.
• Personalized Learning Paths: Recommends what to study next based on a student's progress.
• AI Course Tutor: Answers questions grounded in uploaded learning materials.
• Adaptive Assessments: Generates quizzes to evaluate understanding and track topic-level mastery.

Our goal is to help students understand not only individual concepts but also how an entire course fits together.

How we built it

Course Universe is designed around a pipeline that combines multimodal content processing, AI-powered retrieval, a concept knowledge graph, and learner-progress tracking.

1. Content ingestion: Process course materials such as textbooks, slides, and lecture videos.
2. Concept extraction: Identify key topics, subtopics, and prerequisite relationships.
3. Source grounding: Preserve references to the original pages, slides, and video timestamps.
4. Knowledge visualization: Represent concepts as connected nodes in an interactive learning map.
5. Adaptive learning: Use assessment results to estimate topic mastery and recommend suitable next steps.

The architecture is intended to bring content understanding, trustworthy explanations, and personalized learning into one experience.

Challenges we ran into

One major challenge is extracting meaningful concepts from different formats while preserving accurate source references.

Another challenge is determining which concepts depend on others. A student who struggles with an advanced topic may first need to revisit a fundamental concept.

We also need to balance flexible AI-generated explanations with reliable source grounding, and ensure that quiz results lead to useful rather than misleading learning recommendations.

Accomplishments that we're proud of

We are proud of the idea of transforming a course from a collection of disconnected files into an explorable knowledge universe.

Our key design goals are source-linked explanations, visual concept relationships, personalized learning routes, and a feedback loop that connects assessment results to future study recommendations.

Together, these ideas provide a foundation for a more connected and student-centered learning experience.

What we learned

Developing Course Universe reinforced the importance of trustworthy AI in education. A useful learning assistant should do more than generate convincing answers: it should show where its information comes from and distinguish supported answers from information that is not covered by the course material.

We also learned that personalization requires tracking understanding over time, identifying prerequisite gaps, and evaluating whether recommendations actually help students improve.

What's next for Course Universe

Our next steps are to develop and refine the interactive knowledge map, improve multimodal extraction, strengthen source-level citations, and implement more robust learner modeling.

We also aim to introduce misconception-aware tutoring, spaced-repetition recommendations, voice-based learning, and multilingual explanations.

Finally, we plan to evaluate citation accuracy, retrieval quality, question correctness, and learning progress across simulated student profiles.

Our vision: Make every course an interactive universe of connected knowledge, where every student can discover what they know, understand what they are missing, and find their next step toward mastery.

Built With

• generative-ai
• knowledge-graph
• language
• large-language-models
• machine-learning
• natural
• python
• rag
• streamlit