Windsock is Dancing: Adaptive Multimodal Retrieval-Augmented Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhao, Shu, Shen, Tianyi, Ahuja, Nilesh, Tickoo, Omesh, Narayanan, Vijaykrishnan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914116100161536
author Zhao, Shu
Shen, Tianyi
Ahuja, Nilesh
Tickoo, Omesh
Narayanan, Vijaykrishnan
author_facet Zhao, Shu
Shen, Tianyi
Ahuja, Nilesh
Tickoo, Omesh
Narayanan, Vijaykrishnan
contents Multimodal Retrieval-Augmented Generation (MRAG) has emerged as a promising method to generate factual and up-to-date responses of Multimodal Large Language Models (MLLMs) by incorporating non-parametric knowledge from external knowledge bases. However, existing MRAG approaches suffer from static retrieval strategies, inflexible modality selection, and suboptimal utilization of retrieved information, leading to three critical challenges: determining when to retrieve, what modality to incorporate, and how to utilize retrieved information effectively. To address these challenges, we introduce Windsock, a query-dependent module making decisions on retrieval necessity and modality selection, effectively reducing computational overhead and improving response quality. Additionally, we propose Dynamic Noise-Resistance (DANCE) Instruction Tuning, an adaptive training strategy that enhances MLLMs' ability to utilize retrieved information while maintaining robustness against noise. Moreover, we adopt a self-assessment approach leveraging knowledge within MLLMs to convert question-answering datasets to MRAG training datasets. Extensive experiments demonstrate that our proposed method significantly improves the generation quality by 17.07% while reducing 8.95% retrieval times.
format Preprint
id arxiv_https___arxiv_org_abs_2510_22694
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Windsock is Dancing: Adaptive Multimodal Retrieval-Augmented Generation
Zhao, Shu
Shen, Tianyi
Ahuja, Nilesh
Tickoo, Omesh
Narayanan, Vijaykrishnan
Computer Vision and Pattern Recognition
Computation and Language
Information Retrieval
Multimodal Retrieval-Augmented Generation (MRAG) has emerged as a promising method to generate factual and up-to-date responses of Multimodal Large Language Models (MLLMs) by incorporating non-parametric knowledge from external knowledge bases. However, existing MRAG approaches suffer from static retrieval strategies, inflexible modality selection, and suboptimal utilization of retrieved information, leading to three critical challenges: determining when to retrieve, what modality to incorporate, and how to utilize retrieved information effectively. To address these challenges, we introduce Windsock, a query-dependent module making decisions on retrieval necessity and modality selection, effectively reducing computational overhead and improving response quality. Additionally, we propose Dynamic Noise-Resistance (DANCE) Instruction Tuning, an adaptive training strategy that enhances MLLMs' ability to utilize retrieved information while maintaining robustness against noise. Moreover, we adopt a self-assessment approach leveraging knowledge within MLLMs to convert question-answering datasets to MRAG training datasets. Extensive experiments demonstrate that our proposed method significantly improves the generation quality by 17.07% while reducing 8.95% retrieval times.
title Windsock is Dancing: Adaptive Multimodal Retrieval-Augmented Generation
topic Computer Vision and Pattern Recognition
Computation and Language
Information Retrieval
url https://arxiv.org/abs/2510.22694