TabRAG: Improving Tabular Document Question Answering for Retrieval Augmented Generation via Structured Representations

2025/11/10 by Jacob Si, Mike Qu, Si, Jacob +5
Computer Science · #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Handwritten Text Recognition Techniques #Information Retrieval (cs.IR) #Machine Learning (cs.LG) #Multimodal Machine Learning Applications #Parsing #Question answering #Segmentation #Semantics (computer science) #Suite #Table (database) #Text segmentation #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2511.06582

published in arXiv (Cornell University) (Cornell University)

openalex publication_date 2025/11/10 · openalex created_date 2025/11/12 · openalex updated_date 2026/08/05

Abstract

Incorporating external knowledge bases in traditional retrieval-augmented generation (RAG) relies on parsing the document, followed by querying a language model with the parsed information via in-context learning. While effective for text-based documents, question answering on tabular documents often fails to generate plausible responses. Standard parsing techniques lose the two-dimensional structural semantics critical for cell interpretation. In this work, we present TabRAG, a parsing-based RAG framework designed to improve tabular document question answering via structured representations. Our framework consists of layout segmentation that decomposes the document inputs into a series of components, enabling fine-grained extraction. Subsequently, a vision language model parses and extracts the document tables into a hierarchically structured representation. In order to cater various table styles and formats, we integrate a self-generated in-context learning module that guides the table extraction process. Experimental results demonstrate that TabRAG outperforms existing popular parsing techniques across a broad suite of evaluation and ablation benchmarks. Code is available at: https://github.com/jacobyhsi/TabRAG.

Citations

• TableRAG: A Retrieval Augmented Generation Framework for Heterogeneous Document Reasoning
• TabRep: Training Tabular Diffusion Models with a Simple and Effective Continuous Representation
• Bridging Relevance and Reasoning: Rationale Distillation in Retrieval-Augmented Generation
• VisRAG: Vision-based Retrieval-augmented Generation on Multi-modality Documents
• StructRAG: Boosting Knowledge Intensive Reasoning of LLMs via Inference-time Hybrid Information Structurization
• InstructRAG: Instructing Retrieval-Augmented Generation via Self-Synthesized Rationales
• InterpreTabNet: Distilling Predictive Signals from Tabular Data by Salient Feature Interpretation
• Why Tabular Foundation Models Should Be a Research Priority
• TableVQA-Bench: A Visual Question Answering Benchmark on Multiple Table Domains
• From Local to Global: A Graph RAG Approach to Query-Focused Summarization
• UniTable: Towards a Unified Framework for Table Recognition via Self-Supervised Pretraining
• A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications
• Chain-of-Table: Evolving Tables in the Reasoning Chain for Table Understanding
• Nougat: Neural Optical Understanding for Academic Documents
• Better Zero-Shot Reasoning with Role-Play Prompting
• Reflexion: Language Agents with Verbal Reinforcement Learning
• Toolformer: Language Models Can Teach Themselves to Use Tools
• When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories
• Hierarchical multimodal transformers for Multi-Page DocVQA
• Distributing Accountability, Not Capability: Phase Separation and the LLM Workflow Quadrant in Autonomous AI Agent Architectures
• TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second
• LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking
• Self-Consistency Improves Chain of Thought Reasoning in Language Models
• ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning
• BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence
• DocFormer: End-to-End Transformer for Document Understanding
• DocVQA: A Dataset for VQA on Document Images
• Language Models are Few-Shot Learners
• Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
• REALM: Retrieval-Augmented Language Model Pre-Training
• PlotQA: Reasoning over Scientific Plots
• TabNet: Attentive Interpretable Tabular Learning
• PubLayNet: largest dataset ever for document layout analysis
• Compositional Semantic Parsing on Semi-Structured Tables

Related

• Beyond Extraction: Contextualising Tabular Data for Efficient Summarisation by Language Models
• TableRAG: Million-Token Table Understanding with Language Models
• Tabular Embedding Model (TEM): Finetuning Embedding Models For Tabular RAG Applications
• TableRAG: A Retrieval Augmented Generation Framework for Heterogeneous Document Reasoning
• FlexRAG: A Flexible and Comprehensive Framework for Retrieval-Augmented Generation