{"sourceUrl":null,"slug":"itb25s1p1-1pdf-43918a","apex":{"label":"AI-driven Interactive Hierarchical Concept Maps","id":"n1","children":[{"text":"Traditional concept maps lack interactivity and scalability, necessitating enhancements with AI capabilities and interactive features in digital learning environments.","children":[{"parentId":"n2","id":"n3","label":"Proposed AI Concept Map System","text":"The proposed system leverages artificial intelligence to generate initial concept maps, refined through human-in-the-loop input for pedagogical accuracy and relevance.","children":[{"children":[],"type":"DETL","text":"A custom-designed interface enables learners to explore concepts at varying levels of depth, promoting active engagement and quick navigation.","label":"Interactive System Features","parentId":"n3","id":"n4"}],"type":"CONC"}],"type":"CONC","parentId":"n1","id":"n2","label":"AI-powered Concept Maps Need"},{"label":"Visualization Role in Learning","parentId":"n1","id":"n5","children":[{"parentId":"n5","id":"n6","label":"Early Visualization Studies","text":"Early work by Naps et al. emphasized visual engagement, particularly in computer science education, contributes significantly to student understanding and motivation.","children":[],"type":"DETL"},{"label":"Navigable Concept Map Systems","id":"n7","parentId":"n5","type":"DETL","children":[],"text":"The CoMPASS project and ActiveMath platform introduced navigable and adaptive visual structures to improve structural understanding and navigation efficiency."},{"type":"JUST","children":[],"text":"Hollingsworth and Narayanan argue that interactive features like concept maps should be standard components of digital textbooks.","label":"Interactive Features Standard","id":"n8","parentId":"n5"},{"label":"Self-Regulated Learning Support","id":"n9","parentId":"n5","type":"DETL","children":[],"text":"Barria-Pineda et al. developed visualization tools supporting self-regulated learning via concept-level mapping."},{"parentId":"n5","id":"n10","label":"Ontology-Oriented Learning Systems","text":"In ontology-oriented learning systems, concept maps serve as structured, navigable knowledge representations, exemplified by TM4L.","children":[],"type":"DETL"},{"text":"Many studies have shown that concept maps improve learning outcomes, particularly for complex material.","type":"FIND","children":[{"label":"Meta-analysis Effectiveness","parentId":"n11","id":"n12","children":[],"type":"STAT","text":"A meta-analysis by Schroeder et al. reviewed 142 studies and found concept maps significantly more effective than reading texts (g = .39) or reviewing outlines (g = .28)."},{"parentId":"n11","id":"n13","label":"Anatomy Course Study","text":"Bolatli and Bolatli reported higher post-test scores and lower cognitive load in anatomy students using predefined concept maps.","children":[],"type":"STAT"},{"label":"Interactive Glossary Engagement","parentId":"n11","id":"n14","children":[],"type":"DETL","text":"Elgendi and Shaffer demonstrated interactive glossary maps increased student engagement and repeated glossary use in a computer science e-textbook."},{"type":"DETL","children":[],"text":"Schwab et al. introduced booc.io, a system with drill-down hierarchical concept maps supporting adaptive navigation and targeted feedback.","label":"Booc.io System Features","id":"n15","parentId":"n11"},{"label":"LLM Concept Map Construction","parentId":"n11","id":"n16","children":[{"type":"STAT","children":[],"text":"Their evaluation of GPT-4o demonstrated strong performance, effectively extracting key concepts and accurately identifying hierarchical and cross-topic connections.","label":"GPT-4o Performance","id":"n17","parentId":"n16"},{"parentId":"n16","id":"n18","label":"LLM Map Structure Insight","text":"LLMs could generate concept maps that differ from textbook structure, reflecting a more logical and content-based organization.","children":[],"type":"INSG"}],"type":"DETL","text":"Ma and Chen proposed an LLM framework for automated concept map construction from e-books, including segmentation, key concept extraction, and relationship identification."},{"text":"Kluga et al. integrated causal concept maps into an intelligent anatomy textbook for personalized navigation, quiz adaptation, and content feedback.","type":"DETL","children":[],"id":"n19","parentId":"n11","label":"Causal Concept Maps"},{"text":"Wehnert et al. presented a dynamic visualization system for exploring concept hierarchies extracted from legal textbooks, supporting various navigation modes.","children":[],"type":"DETL","parentId":"n11","id":"n20","label":"Legal Concept Hierarchy System"}],"id":"n11","parentId":"n5","label":"Concept Maps Improve Learning"}],"type":"CONC","text":"Numerous studies have highlighted the pivotal role of visualization in enhancing the comprehension and retention of educational content."},{"id":"n21","parentId":"n1","label":"Interactive Hierarchical Concept Map Platform","text":"This project explores the implementation and evaluation of an interactive, hierarchical concept map platform designed to enhance digital learning experiences.","type":"CONC","children":[{"text":"The interface introduces a layered, LLM-augmented design supporting on-demand concept generation, infinite drill-down navigation, and in-context information display.","children":[],"type":"INSG","parentId":"n21","id":"n22","label":"Interface Layered LLM Design"},{"table":{"rows":[{"label":"LLM Use","cells":["Automates extraction of concepts and relationships","Employs LLMs to generate concept maps"]},{"label":"Real Educational Context","cells":["Does not address use in real educational contexts","Addresses use with validation based on real student feedback"]},{"label":"Interactive UI/UX Design","cells":["Lacks production-ready interactive UI/UX design","Focuses on production-ready interactive UI/UX design"]},{"label":"Cognitive Overload","cells":["Overlooks cognitive overload from large, flat maps","Addresses via hierarchical drill-down navigation and embedded pedagogical content"]},{"cells":["Uses full course materials","Plans to incorporate RAG in future iterations"],"label":"Content Alignment"}],"cols":["Aspect","Ma and Chen System","Proposed System"]},"text":"Our system, similar to Ma and Chen, uses LLMs for concept map generation but addresses interactive UI/UX design and cognitive overload.","type":"CMPR","children":[],"id":"n23","parentId":"n21","label":"Current System vs Ma and Chen"},{"id":"n24","parentId":"n21","label":"System Applied Courses","text":"The system was applied in Object-Oriented Programming and Data Structures, and UI Design and AI-Assisted Frontend Development university courses.","type":"EXMP","children":[]},{"type":"SUBC","children":[{"parentId":"n25","id":"n26","label":"Main Map Functionality","text":"Each concept node could be expanded to explore subtopics, allowing students to see the “big picture” and access detailed content.","children":[],"type":"DETL"}],"text":"The Main Map of the Course functioned as a central, interactive overview of the entire curriculum, visually organizing key topics and their relationships.","label":"Main Course Map Overview","id":"n25","parentId":"n21"},{"label":"Information Panel Details","parentId":"n21","id":"n27","children":[{"text":"This feature allowed learners to engage with the content without leaving the map interface, maintaining flow and minimizing distractions.","type":"INSG","children":[],"id":"n28","parentId":"n27","label":"Information Panel Benefit"}],"type":"SUBC","text":"The Information Panel provided contextual details for each concept selected within the map, displaying concise explanations, examples, or supporting materials."},{"text":"A Child Concept Map is a secondary, more focused map that expands upon a specific node from the main concept map.","children":[{"label":"Child Map Purpose","id":"n30","parentId":"n29","type":"JUST","children":[],"text":"Child maps support hierarchical learning by allowing users to progressively explore concepts at increasing levels of depth without overwhelming them."}],"type":"SUBC","parentId":"n21","id":"n29","label":"Child Concept Map Expansion"},{"type":"SUBC","children":[{"children":[],"type":"DETL","text":"When a student reaches a terminal node and seeks further explanation, the system can generate an extended concept map based on the topic's semantic description.","label":"Dynamic Sub-map Generation","parentId":"n31","id":"n32"},{"label":"Adaptive Learning Environment","id":"n33","parentId":"n31","type":"INSG","children":[],"text":"This feature transforms the concept map into an adaptive learning environment capable of supporting personalized, self-directed exploration across limitless depth within a subject domain."}],"text":"The Infinite Drill-Down AI-Based Domain Exploration feature enables learners to move beyond predefined content by dynamically generating new sub-maps using AI.","label":"Infinite Drill-Down AI Exploration","id":"n31","parentId":"n21"}]},{"parentId":"n1","id":"n34","label":"AI-Driven Map Generation Process","text":"The system generates interactive concept maps by querying LLMs, such as GPT-4o-mini, based on provided course metadata and prompt-tuning commands.","children":[{"children":[{"text":"The instructor defines a new course by providing a course title, a free-form course description, and optional prompt-tuning instructions.","children":[],"type":"DETL","parentId":"n35","id":"n36","label":"Step 1: Course creation"},{"label":"Step 2: Initial map generation","parentId":"n35","id":"n37","children":[],"type":"DETL","text":"The system queries the LLM to produce a draft map of key concepts and their relationships."},{"text":"If the initial map is unsatisfactory, instructors can clear it and regenerate a new one from scratch.","children":[],"type":"DETL","parentId":"n35","id":"n38","label":"Step 3: Full map regeneration"},{"parentId":"n35","id":"n39","label":"Step 4: Map refinement","text":"Users can ask a secondary LLM (e.g., ChatGPT) to rephrase or restructure parts of the map for better clarity or alignment.","children":[],"type":"DETL"},{"text":"Clicking a concept opens an interactive panel that fetches descriptive content via the LLM.","type":"DETL","children":[],"id":"n40","parentId":"n35","label":"Step 5: Concept information generation"},{"parentId":"n35","id":"n41","label":"Step 6: Concept information regeneration","text":"Instructors can discard and regenerate concept information if needed.","children":[],"type":"DETL"},{"label":"Step 7: Concept information refinement","id":"n42","parentId":"n35","type":"DETL","children":[],"text":"Generated descriptions can be edited or enhanced using a separate LLM interface (e.g., ChatGPT)."},{"text":"Instructors can expand individual concepts by generating subordinate maps, enabling drill-down exploration.","children":[],"type":"DETL","parentId":"n35","id":"n43","label":"Step 8: Nested concept map generation"}],"type":"DETL","text":"The end-to-end concept map development process includes eight distinct steps, from course creation and initial generation to refinement and nested map creation.","label":"Concept Map Development Steps","parentId":"n34","id":"n35"},{"text":"The refinement process requires manual review of each map level and concept description to validate pedagogical relevance and accuracy.","children":[],"type":"INSG","parentId":"n34","id":"n44","label":"Quality and Coherence Assurance"},{"text":"For highly customized courses, the refinement process can be time-consuming; however, AI-assisted generation significantly reduces the baseline effort.","children":[],"type":"JUST","parentId":"n34","id":"n45","label":"Refinement Process Time"}],"type":"CONC"},{"label":"Limitations of Generic LLM Maps","id":"n46","parentId":"n1","type":"CONC","children":[{"text":"Initial concept maps are generated based only on a course title and concise metadata, risking shallow or misaligned content in specialized domains.","children":[{"label":"Future Work RAG","parentId":"n47","id":"n48","children":[],"type":"DETL","text":"Future work will explore integrating retrieval-augmented generation (RAG) techniques to enable concept map generation based directly on actual course content."}],"type":"SUBC","parentId":"n46","id":"n47","label":"Metadata-Based Generation Risks"},{"type":"SUBC","children":[{"label":"Disorientation Risk","parentId":"n49","id":"n50","children":[],"type":"DETL","text":"As users delve deeper into nested submaps, they may lose awareness of their location within the overall concept structure."},{"label":"Topic Drift Risk","parentId":"n49","id":"n51","children":[],"type":"DETL","text":"In the absence of clear semantic boundaries, LLMs may generate tangential or unrelated subtopics, resulting in conceptual divergence."},{"id":"n52","parentId":"n49","label":"Future UX Strategies","text":"Future work will explore introducing depth limits and developing UX strategies for clearer visual cues to help users maintain orientation.","type":"DETL","children":[]}],"text":"Although infinite drill-down capability is innovative, it also poses challenges like disorientation and topic drift.","label":"Infinite Drill-Down Challenges","id":"n49","parentId":"n46"},{"label":"Submap Consistency Issues","id":"n53","parentId":"n46","type":"SUBC","children":[{"id":"n54","parentId":"n53","label":"Algorithmic Consistency Support","text":"Scalable resolution for submap consistency will require algorithmic support and improved tooling across the concept hierarchy.","type":"DETL","children":[]}],"text":"Because submaps are generated independently, a concept node on a higher-level map may not reproduce its child nodes when opened in a dedicated submap."}],"text":"Several limitations emerge from relying solely on large language models (LLMs) without incorporating course-specific materials."},{"label":"AI Concept Map System Architecture","parentId":"n1","id":"n55","children":[{"label":"Python Backend","id":"n56","parentId":"n55","type":"DETL","children":[],"text":"The Python backend handles HTTP requests, interacts with the OpenAI GPT-4o-mini API, and stores course content in a file-based structure."},{"type":"DETL","children":[],"text":"The frontend uses Vis.js to render dynamic, clickable concept maps, enabling content display and drill-down navigation.","label":"Vis.js Frontend","id":"n57","parentId":"n55"},{"id":"n58","parentId":"n55","label":"Prompt-Engineered Map Generation","text":"Concept maps are generated by dynamically constructing and sending prompts to an LLM, such as GPT-4o-mini, driven by frontend input and course metadata.","type":"SUBC","children":[{"parentId":"n58","id":"n59","label":"LLM Prompt Instructions","text":"The prompt instructs the LLM to generate at least 15 nodes, define a single root concept, and create a tree-like hierarchical structure without cycles or dangling nodes.","children":[],"type":"DETL"},{"text":"The prompt also requires formatting output as Vis.js code, labeling all edges with clear semantic relationships, ensuring domain-specific accuracy, and including a concrete output example.","type":"DETL","children":[],"id":"n60","parentId":"n58","label":"Detailed Prompt Requirements"},{"type":"DETL","children":[],"text":"The LLM response is cached as a plain text file, serving subsequent requests directly from the local file system to avoid redundant OpenAI API calls.","label":"Map Caching","id":"n61","parentId":"n58"}]},{"label":"Prompt-Engineered Concept Pages","parentId":"n55","id":"n62","children":[{"label":"HTML Generation Requirements","id":"n63","parentId":"n62","type":"DETL","children":[],"text":"The prompt instructs the model to generate valid HTML fragments using semantic tags, format code snippets with pre/code, and mathematical content with LaTeX/MathJax."}],"type":"SUBC","text":"The system generates individual concept pages as HTML fragments through targeted prompt engineering when a user selects a concept node."},{"parentId":"n55","id":"n64","label":"Content Management Panel","text":"The Content Management Panel is a core instructor-facing tool enabling efficient oversight of all AI-generated concept map content.","children":[{"text":"Instructors can create/manage courses, regenerate/refine maps, and customize individual concept pages using third-party LLMs for better alignment.","type":"DETL","children":[],"id":"n65","parentId":"n64","label":"Instructor Management Capabilities"}],"type":"SUBC"}],"type":"CONC","text":"The system features a lightweight, modular architecture that combines AI-driven content generation with interactive frontend visualization."},{"text":"This section presents the design and results of a student survey evaluating the perceived effectiveness of AI-powered concept map interfaces.","children":[{"text":"Hierarchical concept maps were developed for two courses at American University Kyiv, and a survey was administered to students.","type":"SUBC","children":[{"text":"Concept maps were developed for Object-Oriented Programming and Data Structures for first-year Bachelor of Software Engineering/AI and Data Science students.","type":"EXMP","children":[{"id":"n69","parentId":"n68","label":"Course 1 Content Count","text":"42 concept maps and 117 individual concept pages were generated and integrated into Course 1.","type":"STAT","children":[]},{"label":"Course 1 Survey Participants","parentId":"n68","id":"n70","children":[],"type":"STAT","text":"24 students participated in the survey for Course 1."}],"id":"n68","parentId":"n67","label":"Course 1: OOP & Data Structures"},{"text":"Concept maps were developed for UI Design and AI-Assisted Frontend Development for first-year Bachelor of Software Engineering and AI program students.","children":[{"text":"27 concept maps and 128 concept pages were generated and refined for Course 2.","children":[],"type":"STAT","parentId":"n71","id":"n72","label":"Course 2 Content Count"},{"text":"11 students participated in the survey for Course 2.","type":"STAT","children":[],"id":"n73","parentId":"n71","label":"Course 2 Survey Participants"}],"type":"EXMP","parentId":"n67","id":"n71","label":"Course 2: UI Design & AI Frontend"},{"id":"n74","parentId":"n67","label":"Total Survey Responses","text":"A total of 35 surveys were received from both courses.","type":"STAT","children":[]},{"parentId":"n67","id":"n75","label":"Survey Administration","text":"The survey was proposed at the final part of the course as a means of recalling course content and evaluating the concept map’s effectiveness for structured review.","children":[],"type":"DETL"},{"text":"Questions assessed overall satisfaction, engagement, effectiveness for review/learning, and usefulness of drill-down navigation.","type":"DETL","children":[],"id":"n76","parentId":"n67","label":"Key Survey Questions"}],"id":"n67","parentId":"n66","label":"Survey Design Overview"},{"text":"The analysis shows high student satisfaction and perceived effectiveness, quantitatively confirming that the interactive and visual structure of concept maps was well-received.","type":"SUBC","children":[{"id":"n78","parentId":"n77","label":"Overall Satisfaction Score","text":"Students reported an average satisfaction score of 8.91 out of 10 with their learning experience using the concept map app.","type":"STAT","children":[]},{"label":"Learning Effectiveness Score","id":"n79","parentId":"n77","type":"STAT","children":[],"text":"Students rated the app’s usefulness for learning new content at an average of 8.31 out of 10."},{"label":"Engagement & Enjoyment","parentId":"n77","id":"n80","children":[],"type":"STAT","text":"A large majority (91.4%) of students agreed that the app made learning more engaging or enjoyable."},{"id":"n81","parentId":"n77","label":"Drill-Down Usefulness","text":"100% of students confirmed that the ability to click on concept names and navigate through nested maps was helpful.","type":"STAT","children":[]},{"text":"Students expressed varying preferences for reviewing course material using interactive concept maps versus traditional materials.","type":"DETL","children":[{"type":"STAT","children":[],"text":"60% of students found both interactive concept maps and traditional materials equally effective for review and recall.","label":"Both Equally Effective","id":"n83","parentId":"n82"},{"type":"STAT","children":[],"text":"34.3% of students preferred interactive concept maps for reviewing course material.","label":"Preferred Interactive Maps","id":"n84","parentId":"n82"},{"parentId":"n82","id":"n85","label":"Preferred Text-Based Materials","text":"Only 5.7% of students preferred text-based materials alone for reviewing course material.","children":[],"type":"STAT"}],"id":"n82","parentId":"n77","label":"Preferred Review Methods"},{"label":"Most Useful Features","parentId":"n77","id":"n86","children":[{"type":"DETL","children":[],"text":"11 students mentioned appreciating the ability to click, navigate, and explore deeper levels of topics.","label":"Clickability and Depth Navigation","id":"n87","parentId":"n86"},{"text":"9 students highlighted the clarity and usefulness of the visual concept map structure and topic relationships.","children":[],"type":"DETL","parentId":"n86","id":"n88","label":"Visual Structure and Relationships"}],"type":"DETL","text":"Open-ended responses revealed clickability/depth navigation and visual structure/relationships between topics as major advantages."},{"children":[{"text":"7 students felt the app was already perfect or good enough, stating “everything is perfect” or “all is good”.","type":"STAT","children":[],"id":"n90","parentId":"n89","label":"App Perfect or Good"},{"children":[],"type":"STAT","text":"8 students recommended UI improvements, including design changes for a more pleasant visual side and a dark theme.","label":"UI Improvement Suggestions","parentId":"n89","id":"n91"},{"text":"6 students recommended improvements in current interactions, citing issues with dragging, zooming, and scrolling the graph.","type":"STAT","children":[],"id":"n92","parentId":"n89","label":"Interaction Improvement Suggestions"},{"text":"Various individual ideas were suggested, such as online collaboration, shared comments, quizzes, and the ability to ask questions.","children":[],"type":"DETL","parentId":"n89","id":"n93","label":"Additional Functionality Ideas"}],"type":"DETL","text":"Students suggested improvements related to UI, current interactions, and additional functionalities.","label":"Suggested Improvements","parentId":"n77","id":"n89"},{"text":"High ratings across both courses suggest AI-generated maps are relevant for learning, and the hierarchical drill-down structure offers an effective human-computer interaction pattern.","type":"INSG","children":[],"id":"n94","parentId":"n77","label":"Evaluation Conclusion"}],"id":"n77","parentId":"n66","label":"Survey Results Summary"}],"type":"CONC","parentId":"n1","id":"n66","label":"Evaluation: Student Feedback"},{"text":"This study presented the design and implementation of an AI-driven system for generating interactive hierarchical concept maps for digital learning environments and intelligent textbooks.","children":[{"label":"System as Intelligent Textbook","id":"n96","parentId":"n95","type":"INSG","children":[],"text":"The system positions itself as a practical implementation of the intelligent textbook paradigm, augmenting content delivery and navigational structure for student-centered learning."},{"label":"Evaluation Results Summary","id":"n97","parentId":"n95","type":"DETL","children":[],"text":"Evaluation results, based on 69 AI-generated concept maps and 245 unique concept descriptions across two undergraduate computer science courses, confirm the system's educational value."},{"label":"Key Findings Reinforcement","id":"n98","parentId":"n95","type":"INSG","children":[],"text":"Feedback from 35 student surveys underscored the potential of integrating AI-generated maps into intelligent textbooks and online courses to enhance comprehension, engagement, and structured review."},{"parentId":"n95","id":"n99","label":"Future Work Directions","text":"Future work will address current limitations identified in the system’s architecture and usage, focusing on several key directions.","children":[{"id":"n100","parentId":"n99","label":"RAG Integration for Content","text":"One key direction is the integration of retrieval-augmented generation (RAG) techniques to allow the LLM to incorporate course-specific materials into the generation process.","type":"DETL","children":[]},{"parentId":"n99","id":"n101","label":"Usability Enhancements","text":"To improve usability and mitigate disorientation, future enhancements will include research on depth-limiting mechanisms and UX improvements for visual context and navigational cues.","children":[],"type":"DETL"},{"text":"Algorithmic methods and instructor-facing tools will be developed to support structural validation and map coherence across different levels of hierarchy.","children":[],"type":"DETL","parentId":"n99","id":"n102","label":"Submap Consistency Resolution"},{"label":"Refinement Process Integration","id":"n103","parentId":"n99","type":"DETL","children":[],"text":"The refinement process will be more tightly integrated into the instructor-facing UI for seamless editing and regeneration of maps and concept pages within the same workflow."},{"label":"Admin Dashboard Extension","parentId":"n99","id":"n104","children":[],"type":"DETL","text":"The admin dashboard will be extended to provide broader control and monitoring of content quality."},{"type":"DETL","children":[],"text":"Additional research will explore the system’s effectiveness in other domains such as mathematics, management, and postgraduate software engineering education.","label":"Other Domain Exploration","id":"n105","parentId":"n99"},{"id":"n106","parentId":"n99","label":"Agentic AI Workflows","text":"The implementation of agentic AI workflows, where multiple autonomous agents coordinate multi-step concept map construction and refinement, will be explored.","type":"DETL","children":[]}],"type":"SUBC"}],"type":"CONC","parentId":"n1","id":"n95","label":"Conclusion and Future Work"},{"children":[],"type":"DETL","text":"The author thanks American University Kyiv instructors Roman Tymoshuk, Andrii Tsabanov, and Ivan Danilov for reviewing and integrating the concept maps into the learning process.","label":"Acknowledgements","parentId":"n1","id":"n107"},{"text":"The author used ChatGPT-4 and Grammarly to check grammar, spelling, improve writing style, paraphrase, and reword, taking full responsibility for the content.","children":[],"type":"DETL","parentId":"n1","id":"n108","label":"Generative AI Declaration"}],"type":"APEX","text":"This paper explores the design and implementation of AI-driven interactive concept maps as components of intelligent textbooks and digital learning environments."},"sourceType":"file","contentType":"explainer","sharedAt":{"_seconds":1780225254,"_nanoseconds":403000000},"title":"AI-driven Interactive Hierarchical Concept Maps"}