Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

MultiFair: Multimodal Balanced Fairness-Aware Medical Classification with Dual-Level Gradient Modulation.

Md Zubair, Hao Zheng, Grayson W Armstrong

IEEE Transactions on Medical Imaging

|July 15, 2026

Summary

This study introduces MultiFair, a new method for multimodal medical classification that tackles uneven data learning and demographic bias. MultiFair ensures fairer and more balanced AI diagnostic models by modulating training gradients.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

A nationwide retrospective study of toy-related ocular injuries in children presenting to emergency departments in the United States.

Journal of AAPOS : the official publication of the American Association for Pediatric Ophthalmology and Strabismus·2026

Same author

Cataract Surgery and the Risk of Conversion from Dry to Neovascular Age-related Macular Degeneration in the IRIS© Registry.

Ophthalmology·2026

Same author

Seasonal Pattern of Glaucoma Emergencies in New England.

Ophthalmology. Glaucoma·2026

Same author

Fall-related open globe injuries are associated with systemic health changes.

Eye (London, England)·2026

Same author

A 19-Year-Old Man with Loss of Vision.

NEJM evidence·2025

Same author

Ocular trauma in microgravity: In-flight diagnostics and extraterrestrial strategies for management.

Survey of ophthalmology·2025

Area of Science:

Artificial Intelligence
Medical Informatics
Machine Learning

Background:

Multimodal learning models are crucial for reliable medical diagnosis.
Existing models struggle with uneven modality learning and demographic bias, leading to unfair outcomes.
These issues can be interconnected, with modalities favoring specific groups.

Purpose of the Study:

To propose a novel approach, MultiFair, for fair and balanced multimodal medical classification.
To address the challenges of uneven modality learning and demographic bias in AI-driven medical decisions.
To improve the reliability and unbiased nature of AI diagnostic systems.

Main Methods:

Developed MultiFair, a novel approach for multimodal medical classification.

Related Experiment Videos

Implemented a dual-level gradient modulation process to dynamically adjust training gradients.

Modulated gradients at both data modality and demographic group levels.

Evaluated on three real-world medical datasets with diverse demographic attributes and missing modalities.

Main Results:

MultiFair effectively addresses uneven modality learning and demographic bias.
Experimental results demonstrate the effectiveness of the proposed dual-level gradient modulation.
The approach shows promise in both multiclass classification and missing-modality scenarios.
Achieved fairer and more balanced performance across different demographic groups.

Conclusions:

MultiFair offers a significant advancement in developing fair and unbiased multimodal medical AI.
The dual-level gradient modulation is key to mitigating imbalanced and unfair learning.
This method enhances the trustworthiness of AI in medical decision-making.