mlsanityData to 7 Oct 2026

FLOWREADER: Min-Cost Flow Optimization for Multi-Modal Long Document Q&A

arXiv 2606.07235 ↗First posted 5 Jun 2026cs.IR cs.LGcited by 1

On the map: Retrieval-augmented generation → Multimodal RAG

Abstract (arXiv)

Answering questions about long multimodal documents requires distributing a fixed evidence budget across relevant facets in text, tables, figures, and slides while avoiding near-duplicates. We present flowreader, which formulates evidence selection as a single minimum-cost flow problem with capacity limits over a multimodal content graph. Spectral decomposition identifies latent aspects of query-relevant content and allocates the budget among them in proportion to their spectral energy. These capacity limits enforce aspect coverage during routing without requiring a language-model planning call. Query-conditioned costs prioritize chains of relevant, mutually consistent evidence. Decomposing the optimal flow produces short evidence chains, which a vision-language model reads in parallel and a reasoner reconciles. On VisDoMBench with Qwen3-VL-32B, flowreader achieves the highest macro accuracy (68.9), surpassing the strongest prior system by 2.7 points, leading on three of five subsets and attaining the highest worst-subset accuracy. It uses a measured 17.5 content nodes per query and maintains its lead at 12.9. Ablation studies with a fixed graph, scorer, reader, and judge show that cost design drives accuracy, capacity limits preserve it while using about three-quarters of the reader tokens required by shortest-path routing without these limits on the same network, and spectral aspects align with LLM-generated sub-questions without a planning call.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Multimodal RAG, one of 15 directions in Retrieval-augmented generation. Builds on 49 of its 73 references that are in the map. First posted on arXiv 5 Jun 2026.

Builds on

49 of 73 references.

Closest ancestors · 6

Same direction, newest first

  1. 2025Scaling Beyond Context: A Survey of Multimodal Retrieval-Augmented Generation for Document Understanding17 Oct
  2. 2025RAG-Anything: All-in-One RAG Framework14 Oct
  3. 2025A Multi-Granularity Retrieval Framework for Visually-Rich Documents1 May
  4. 2025ViDoRAG: Visual Document Retrieval-Augmented Generation via Dynamic Iterative Reasoning Agents25 Feb
  5. 2024VisDoM: Multi-Document QA with Visually Rich Elements Using Multimodal Retrieval-Augmented Generation14 Dec
  6. 2024VisRAG: Vision-based Retrieval-augmented Generation on Multi-modality Documents14 Oct

Other references · 43

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,383
  4. 2025Qwen3-VL Technical Reportcited by 2,534
  5. 2013Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distancescited by 2,277
  6. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,800
  7. 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,411
  8. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,054
  9. 2025A-MEM: Agentic Memory for LLM Agentscited by 1,000
  10. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questionscited by 911
  11. 2025Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memorycited by 722
  12. 2023Active Retrieval Augmented Generationcited by 506
31 more
  1. 2021A Dataset of Information-Seeking Questions and Answers Anchored in Research Paperscited by 475
  2. 2024Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexitycited by 437
  3. 2025Zep: A Temporal Knowledge Graph Architecture for Agent Memorycited by 372
  4. 2023Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergycited by 343
  5. 2024LightRAG: Simple and Fast Retrieval-Augmented Generationcited by 310
  6. 2024Nomic Embed: Training a Reproducible Long Context Text Embeddercited by 255
  7. 2024MinerU: An Open-Source Solution for Precise Document Content Extractioncited by 255
  8. 2024ColPali: Efficient Document Retrieval with Vision Language Modelscited by 190
  9. 2023SlideVQA: A Dataset for Document Visual Question Answering on Multiple Imagescited by 175
  10. 2021FeTaQA: Free-form Table Question Answeringcited by 174
  11. 2024MMLongBench-Doc: Benchmarking Long-context Document Understanding with Visualizationscited by 174
  12. 2024Speculative RAG: Enhancing Retrieval Augmented Generation through Draftingcited by 76
  13. 2024SPIQA: A Dataset for Multimodal Question Answering on Scientific Paperscited by 75
  14. 2025PathRAG: Pruning Graph-based Retrieval Augmented Generation with Relational Pathscited by 60
  15. 2024GraphReader: Building Graph-based Agent to Enhance Long-Context Abilities of Large Language Modelscited by 42
  16. 2024Auto-RAG: Autonomous Retrieval-Augmented Generation for Large Language Modelscited by 41
  17. 2024UDA: A Benchmark Suite for Retrieval Augmented Generation in Real-world Document Analysiscited by 40
  18. 2024Stochastic RAG: End-to-End Retrieval-Augmented Generation through Expected Utility Maximizationcited by 29
  19. 2025MA-RAG: Multi-Agent Retrieval-Augmented Generation via Collaborative Chain-of-Thought Reasoningcited by 24
  20. 2019Improving the Similarity Measure of Determinantal Point Processes for Extractive Multi-Document Summarizationcited by 22
  21. 2025DualRAG: A Dual-Process Approach to Integrate Reasoning and Retrieval for Multi-Hop Question Answeringcited by 20
  22. 2025Reasoning RAG via System 1 or System 2: A Survey on Reasoning Agentic Retrieval-Augmented Generation for Industry Challengescited by 19
  23. 2025MMGraphRAG: Bridging Vision and Language with Interpretable Multimodal Knowledge Graphscited by 16
  24. 2025Cog-RAG: Cognitive-Inspired Dual-Hypergraph with Theme Alignment Retrieval-Augmented Generationcited by 9
  25. 2022OTExtSum: Extractive Text Summarisation with Optimal Transportcited by 8
  26. 2026Query-Aware Flow Diffusion for Graph-Based RAG with Retrieval Guaranteescited by 7
  27. 2024SMART-RAG: Selection using Determinantal Matrices for Augmented Retrievalcited by 6
  28. 2025MOTOR: Multimodal Optimal Transport via Grounded Retrieval in Medical Visual Question Answeringcited by 5
  29. 2025From Ranking to Selection: A Simple but Efficient Dynamic Passage Selector for Retrieval Augmented Generationcited by 4
  30. 2026When Knowledge Is Not Free: Cost-Aware Evidence Selection in Retrieval-Augmented Generationcited by 3
  31. 2025Question-Adaptive Graph Learning for Multi-hop Retrieval Augmented Generationcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026MAGE-RAG: Multigranular Adaptive Graph Evidence for Agentic Multimodal RAG in Long-Document QAcited by 2

Similar, not cited

Close in wording, no citation either way

  1. 2026G2-Reader: Dual Evolving Graphs for Multimodal Document QA
  2. 2026TAP-RAG: Task-Aware Policy Control for Long-Document Multimodal Question Answering
  3. 2026TopoGraphRAG-Bench: Evaluating Multimodal GraphRAG on Layout-Grounded Evidence Reasoning
  4. 2026DocNavRAG: Document-Structured Graph RAG with Stateful Evidence Construction for Complex Document Question Answering
  5. 2026MLDocRAG: Multimodal Long-Context Document Retrieval Augmented Generation
  6. 2026N2N-GQA: Noise-to-Narrative for Graph-Based Table-Text Question Answering Using LLMs
  7. 2026GraphLoom: Reliability-Calibrated Graph Evidence Routing for Multimodal KG-RAG
  8. 2026HKVM-RAG: Key-Value-Separated Hypergraph Evidence Organization for Multi-Hop RAG