A Survey of Multimodal Retrieval-Augmented Generation

Lang Mei, Siyu Mo, Zhihan Yang, Chong Chen
3/26/2025
cs.IRcs.AIcs.CL

No official code found. Best implementation: llm-lab-org/Multimodal-RAG-Survey (531 stars, last commit 2026-02-20). 1 more below.

Abstract

Multimodal Retrieval-Augmented Generation (MRAG) enhances large language models (LLMs) by integrating multimodal data (text, images, videos) into retrieval and generation processes, overcoming the limitations of text-only Retrieval-Augmented Generation (RAG). While RAG improves response accuracy by incorporating external textual knowledge, MRAG extends this framework to include multimodal retrieval and generation, leveraging contextual information from diverse data types. This approach reduces hallucinations and enhances question-answering systems by grounding responses in factual, multimodal knowledge. Recent studies show MRAG outperforms traditional RAG, especially in scenarios requiring both visual and textual understanding. This survey reviews MRAG's essential components, datasets, evaluation methods, and limitations, providing insights into its construction and improvement. It also identifies challenges and future research directions, highlighting MRAG's potential to revolutionize multimodal information retrieval and generation. By offering a comprehensive perspective, this work encourages further exploration into this promising paradigm.

View on arXivView PDF

Code Implementations(2)

llm-lab-org/Multimodal-RAG-SurveyCommunity77%

A Survey on Multimodal Retrieval-Augmented Generation

53129Feb 11, 20258 months ago
multimodal-learningragretrieval-augmented-generation
PanguIR/MRAGSurveyCommunity72%

A Survey of Multimodal Retrieval-Augmented Generation

202Mar 14, 202511 months ago
large-language-modelsllmsmllmsmultimodal-generationmultimodal-large-language-models+3 more

Similar Papers

Ask in Any Modality: A Comprehensive Survey on Multimodal Retrieval-Augmented Generation

94% match1 repo5312025

MRAMG-Bench: A Comprehensive Benchmark for Advancing Multimodal Retrieval-Augmented Multimodal Generation

91% match1 repo192025

Searching for Best Practices in Retrieval-Augmented Generation

90% match1 repo3472024

A Comprehensive Survey of Retrieval-Augmented Generation (RAG): Evolution, Current Landscape and Future Directions

89% match2024

Re-ranking the Context for Multimodal Retrieval Augmented Generation

89% match1 repo5312025

Beyond Text: Optimizing RAG with Multimodal Inputs for Industrial Applications

88% match1 repo722024

Cite this paper

@article{mei2025survey,
  title  = {A Survey of Multimodal Retrieval-Augmented Generation},
  author = {Lang Mei and Siyu Mo and Zhihan Yang and Chong Chen},
  year   = {2025},
  eprint = {2504.08748},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2504.08748v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2504.08748.svg)](https://codeofpaper.com/paper/2504.08748)

Discussion