MEXA: Towards General Multimodal Reasoning with Dynamic Multi-Expert Aggregation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yu, Shoubin, Zhang, Yue, Wang, Ziyang, Yoon, Jaehong, Bansal, Mohit
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909869073760256
author Yu, Shoubin
Zhang, Yue
Wang, Ziyang
Yoon, Jaehong
Bansal, Mohit
author_facet Yu, Shoubin
Zhang, Yue
Wang, Ziyang
Yoon, Jaehong
Bansal, Mohit
contents Combining pre-trained expert models offers substantial potential for scalable multimodal reasoning, but building a unified framework remains challenging due to the increasing diversity of input modalities and task complexity. For instance, medical diagnosis requires precise reasoning over structured clinical tables, while financial forecasting depends on interpreting plot-based data to make informed predictions. To tackle this challenge, we introduce MEXA, a training-free framework that performs modality- and task-aware aggregation of multiple expert models to enable effective multimodal reasoning across diverse and distinct domains. MEXA dynamically selects expert models based on the input modality and the task-specific reasoning demands (i.e., skills). Each expert model, specialized in a modality task pair, generates interpretable textual reasoning outputs. MEXA then aggregates and reasons over these outputs using a Large Reasoning Model (LRM) to produce the final answer. This modular design allows flexible and transparent multimodal reasoning across diverse domains without additional training overhead. We extensively evaluate our approach on diverse multimodal benchmarks, including Video Reasoning, Audio Reasoning, 3D Understanding, and Medical QA. MEXA consistently delivers performance improvements over strong multimodal baselines, highlighting the effectiveness and broad applicability of our expert-driven selection and aggregation in diverse multimodal reasoning tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2506_17113
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MEXA: Towards General Multimodal Reasoning with Dynamic Multi-Expert Aggregation
Yu, Shoubin
Zhang, Yue
Wang, Ziyang
Yoon, Jaehong
Bansal, Mohit
Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Combining pre-trained expert models offers substantial potential for scalable multimodal reasoning, but building a unified framework remains challenging due to the increasing diversity of input modalities and task complexity. For instance, medical diagnosis requires precise reasoning over structured clinical tables, while financial forecasting depends on interpreting plot-based data to make informed predictions. To tackle this challenge, we introduce MEXA, a training-free framework that performs modality- and task-aware aggregation of multiple expert models to enable effective multimodal reasoning across diverse and distinct domains. MEXA dynamically selects expert models based on the input modality and the task-specific reasoning demands (i.e., skills). Each expert model, specialized in a modality task pair, generates interpretable textual reasoning outputs. MEXA then aggregates and reasons over these outputs using a Large Reasoning Model (LRM) to produce the final answer. This modular design allows flexible and transparent multimodal reasoning across diverse domains without additional training overhead. We extensively evaluate our approach on diverse multimodal benchmarks, including Video Reasoning, Audio Reasoning, 3D Understanding, and Medical QA. MEXA consistently delivers performance improvements over strong multimodal baselines, highlighting the effectiveness and broad applicability of our expert-driven selection and aggregation in diverse multimodal reasoning tasks.
title MEXA: Towards General Multimodal Reasoning with Dynamic Multi-Expert Aggregation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2506.17113