# Navigating Large-Scale Document Collections: MuDABench for Multi-Document Analytical QA

arXiv: https://arxiv.org/abs/2604.22239 · HTML: https://mlsanity.com/paper/2604.22239/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 33 references that are in the map.
- First posted on arXiv 24 Apr 2026.

## Builds on (most cited first, 18 of 18)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding](https://mlsanity.com/paper/2308.14508/) (2023)
- [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://mlsanity.com/paper/2404.06654/) (2024)
- [DocVQA: A Dataset for VQA on Document Images](https://mlsanity.com/paper/2007.00398/) (2020)
- [Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Steps](https://mlsanity.com/paper/2011.01060/) (2020)
- [FigureQA: An Annotated Figure Dataset for Visual Reasoning](https://mlsanity.com/paper/1710.07300/) (2017)
- [FinanceBench: A New Benchmark for Financial Question Answering](https://mlsanity.com/paper/2311.11944/) (2023)
- [TableBench: A Comprehensive and Complex Benchmark for Table Question Answering](https://mlsanity.com/paper/2408.09174/) (2024)
- [LongDocURL: a Comprehensive Multimodal Long Document Benchmark Integrating Understanding, Reasoning, and Locating](https://mlsanity.com/paper/2412.18424/) (2024)
- [DocETL: Agentic Query Rewriting and Evaluation for Complex Document Processing](https://mlsanity.com/paper/2410.12189/) (2024)
- [UDA: A Benchmark Suite for Retrieval Augmented Generation in Real-world Document Analysis](https://mlsanity.com/paper/2406.15187/) (2024)
- [FinAgentBench: A Benchmark Dataset for Agentic Retrieval in Financial Question Answering](https://mlsanity.com/paper/2508.14052/) (2025)
- [More Documents, Same Length: Isolating the Challenge of Multiple Documents in RAG](https://mlsanity.com/paper/2503.04388/) (2025)
- [Uncovering Limitations of Large Language Models in Information Seeking from Tables](https://mlsanity.com/paper/2406.04113/) (2024)
- [DeepRead: Document Structure-Aware Reasoning to Enhance Agentic Search](https://mlsanity.com/paper/2602.05014/) (2026)
- [ESG Rating Disagreement and Corporate Total Factor Productivity:Inference and Prediction](https://mlsanity.com/paper/2408.13895/) (2024)

## Built on by (most cited first, 2 of 2)

- [ENTLORE: A Graph-Grounded Benchmark for Latent Organizational Reasoning in Enterprise Question Answering](https://mlsanity.com/paper/2608.10679/) (2026)
- [Towards Reliable AI Data Scientists: Data Agents with Workflow Harnesses](https://mlsanity.com/paper/2609.35255/) (2026)

## Abstract (arXiv)

This paper introduces the task of analytical question answering over large, semi-structured document collections. We present MuDABench, a benchmark for multi-document analytical QA, where questions require extracting and synthesizing information across numerous documents to perform quantitative analysis. Unlike existing multi-document QA benchmarks that typically require information from only a few documents with limited cross-document reasoning, MuDABench demands extensive inter-document analysis and aggregation. Constructed via distant supervision by leveraging document-level metadata and annotated financial databases, MuDABench comprises over 80,000 pages and 332 analytical QA instances. We also propose an evaluation protocol that measures final answer accuracy and uses intermediate-fact coverage as an auxiliary diagnostic signal for the reasoning process. Experiments reveal that standard RAG systems, which treat all documents as a flat retrieval pool, perform poorly. To address these limitations, we propose a multi-agent workflow that orchestrates planning, extraction, and code generation modules. While this approach substantially improves both process and outcome metrics, a significant gap remains compared to human expert performance. Our analysis identifies two primary bottlenecks: single-document information extraction accuracy and insufficient domain-specific knowledge in current systems. MuDABench is available at https://github.com/Zhanli-Li/MuDABench.