Agentic AI Quiz-Based Learning System: Enhancing MCQ Generation via Long-Context Cached Retrieval-Augmented Generation
Abstract
This research-to-practice full paper investigates how Large Language Models (LLMs) with long-context window and enhanced retrieval efficiency can generate context-specific quizzes to address high attrition in engineering education. We aim to enable LLMs to process large, multidisciplinary Artificial Intelligence (AI) datasets, covering topics like Machine Learning, Generative AI, and Neural Networks, from foundational to advanced concepts. A systematic literature review identified gaps in Retrieval-Augmented Generation (RAG) systems, which often retrieve irrelevant chunks due to context limitations, leading to inaccurate or hallucinated responses [1], [2]. Traditional quiz generation lacks modular design, limiting scalability and interpretability. To address this, we developed an agentic long-context RAG architecture using Gemini 1.5's one-million-token window, integrating retrieval, reasoning, and evaluation in a unified pipeline. Our methodology employed a modular Agentic AI system. A Parsing Agent extracts text from academic sources, followed by a Chunking & Storage Agent segmenting content with character overlaps. An Embedding & Indexing Agent generates and indexes vector embeddings, verified by a Verification Agent for topical alignment. For quiz generation, a Retriever Agent uses cosine similarity and multilingual re-ranking, a Selector Agent filters meaningful chunks, a Response Agent leverages cached ground-truth MCQs with an LLM, and an Evaluator Agent assesses…
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.