MultiPDF RAG: An Intelligent Retrieval-Augmented Generation Framework for Multi-Document Question Answering

This paper presented a MultiPDF Retrieval-Augmented Generation (MultiPDF RAG) system designed to enable accurate, scalable, and context-aware question answering across multiple PDF documents. The proposed approach effectively addresses the limitations of traditional keyword-based search systems and standalone large language models by integrating semantic retrieval with generative AI.   By converting document content into semantic embeddings and retrieving only the most relevant document segments, the system ensures that generated responses are grounded in authentic source data. This significantly reduces hallucinations and improves the reliability of answers. The modular architecture allows efficient handling of large document collections while supporting scalability and cloud-based deployment.   The system demonstrates strong potential in domains such as academic research, enterprise knowledge management, legal analysis, technical documentation, and healthcare information systems. Its ability to synthesize information across multiple documents enhances decision-making, productivity, and knowledge accessibility.   Overall, the MultiPDF RAG framework provides a robust and practical solution for intelligent document understanding. With future enhancements in multimodal processing, domain-specific optimization, and source attribution, the system can further evolve into a comprehensive and trustworthy document intelligence platform suitable for real-world applications.

·        A. Omar et al., “Multi-Level Information Retrieval Augmented Generation for Visual QA,” EMNLP, 2024. ACL Anthology   ·        Z. Wang et al., “Document Segmentation Matters for Retrieval-Augmented Generation,” Findings of ACL, 2025. ACL Anthology   ·        M. Kwon et al., “A Dynamic-Selection-Based Retrieval-Augmented Generation Framework: Enhancing Multi-Document QA,” Electronics, 2025. MDPI+1   ·        Z. Liu et al., “Benchmarking Retrieval-Augmented Generation in Multi-Document Settings,” ACM, 2025. ACM Digital Library   ·        T. T. U. Hoang & V. A. Nguyen, “PDF Retrieval Augmented Question Answering,” arXiv, 2025. arXiv   ·        M. Suri et al., “VisDoM: Multi-Document QA with Visually Rich Elements Using Multimodal RAG,” arXiv, 2024. arXiv ·        I. Besrour et al., “SQuAI: Scientific Question-Answering with Multi-Agent RAG,” arXiv, 2025. arXiv   ·        “A Systematic Review of Key Retrieval-Augmented Generation Components,” arXiv, 2025. arXiv   ·        M. Abo El-Enen et al., “A Survey on Retrieval-Augmented Generation (RAG) Models,” Springer, 2025. Springer Link