ABEL: Artificial Buddy for Effective Learning
T. Y. Emmy Lai, Diego Collarana, Héctor Allende-Cid, Tobias Lang, Marlena Flüh, Dena Baghery, Ann-Kathrin Bernards
Lecture notes in computer science
Problems Identified (4)
Educational Content Retrieval: The paper addresses the need to enhance and support education in Data Science and Artificial Intelligence with relevant educational content.
Grounded Explainable Chatbot Responses: The paper addresses the need for chatbot responses that are contextually grounded, personalized, specific, explainable, traceable, and correct.
Educational Content Retrieval: The paper addresses the need to enhance and support education in Data Science and Artificial Intelligence with relevant educational content.
Grounded Explainable Chatbot Responses: The paper addresses the need for chatbot responses that are contextually grounded, personalized, specific, explainable, traceable, and correct.
Proposed Solutions (5)
KG-Driven Educational Chatbot: ABEL is a modular chatbot driven by a knowledge graph for education in Data Science and Artificial Intelligence.
Hybrid KG-RAG Retrieval: ABEL uses a hybrid retrieval architecture combining a dynamic Knowledge Graph with a Retrieval-Augmented Generation pipeline.
Graph And Embedding Retrieval: The system retrieves semantically relevant educational content using multi-hop graph queries and embedding-based similarity search over a curated-resource knowledge graph.
FAQ-Based RAG: ABEL also includes a FAQ-based RAG approach to provide flexible access to learning content.
KG-Driven Educational Chatbot: ABEL is a modular chatbot driven by a knowledge graph for education in Data Science and Artificial Intelligence.
Results (3)
Improved Relevance And Adaptability:
Retrieval And User Evaluation:
Contextual Grounding And Explainable Responses:
Research Domain
AI-supported education and educational chatbots