Adaptive multi-agent AI tutoring platform for data structures and algorithms in undergraduate education

Abstract

This project addresses the persistent challenge faced by undergraduate students in learning data structures and algorithms, particularly regarding conceptual difficulty, limited student engagement, and gaps in traditional teaching methods. To tackle these issues, an artificial intelligence (AI)-powered educational platform named AlgoGPT has been developed, focusing specifically on enhancing students' proficiency in Python programming within the context of data structures and algorithms. AlgoGPT leverages advanced large language models (LLMs), specifically GPT-4o, enhanced through retrieval-augmented generation (RAG). RAG integrates the generative strengths of LLMs with a curated knowledge repository of structured questions and detailed solutions. This enables AlgoGPT to provide highly personalised, context-aware, and accurate educational guidance and feedback.

Key features of AlgoGPT designed to enhance comprehension, engagement, and practical skills include:

• A personalised learning roadmap that dynamically adapts to each student's progress. • Automatic difficulty adjustment mechanisms to ensure learners are appropriately challenged without becoming overwhelmed. • Interactive tools to actively engage students in practising and mastering core algorithmic concepts. • Conceptual explanations combined with step-by-step guidance and clear visual representations to improve understanding of complex ideas and solutions.

By implementing these features, AlgoGPT aims to bridge the gap between theoretical concepts and practical applications. Thereby offering students a comprehensive and adaptive learning experience in data structures and algorithms.