MultiDocFusion: Hierarchical and Multimodal Chunking Pipeline for Enhanced RAG on Long Industrial Documents

2026-04-14

AI Summary

This work addresses the limitations of conventional text chunking methods, which overlook the intricate hierarchical structures of industrial documents, leading to information loss and degraded question-answering performance. To overcome this, the authors propose a multimodal structure-aware chunking approach that uniquely integrates visual document parsing, optical character recognition (OCR), and large language models (LLMs). Their method employs an LLM-driven Document Structure Hierarchical Parsing (DSHP-LLM) module to reconstruct document hierarchies, followed by depth-first search–guided semantic chunking within a retrieval-augmented generation (RAG) framework. By explicitly modeling the multimodal and hierarchical nature of industrial documents, the proposed approach achieves significant improvements over existing baselines, yielding 8–15% gains in retrieval accuracy and 2–3% higher ANLS scores on industrial benchmarks.

Technology Category

• Natural Language Processing: Syntax — Tagging, Chunking & Parsing
• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

RAG-based QA has emerged as a powerful method for processing long industrial documents. However, conventional text chunking approaches often neglect complex and long industrial document structures, causing information loss and reduced answer quality. To address this, we introduce MultiDocFusion, a multimodal chunking pipeline that integrates: (i) detection of document regions using vision-based document parsing, (ii) text extraction from these regions via OCR, (iii) reconstruction of document structure into a hierarchical tree using large language model (LLM)-based document section hierarchical parsing (DSHP-LLM), and (iv) construction of hierarchical chunks through DFS-based grouping. Extensive experiments across industrial benchmarks demonstrate that MultiDocFusion improves retrieval precision by 8-15% and ANLS QA scores by 2-3% compared to baselines, emphasizing the critical role of explicitly leveraging document hierarchy for multimodal document-based QA. These significant performance gains underscore the necessity of structure-aware chunking in enhancing the fidelity of RAG-based QA systems.

Problem

• RAG
• industrial documents
• document structure
• text chunking
• multimodal QA

Innovation

• Multimodal Chunking
• Hierarchical Document Parsing
• RAG
• Vision-based Document Analysis
• LLM-based Structuring