MedMKEB: A Comprehensive Knowledge Editing Benchmark for Medical Multimodal Large Language Models
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (40) · 34106-34114
DOI: 10.1609/aaai.v40i40.40705 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Dexuan Xu und 6 weiteren Autor:innen.
Abstract (Originalsprache)
Recent advances in multimodal large language models (MLLMs) have significantly improved medical AI, enabling it to unify the understanding of visual and textual information. However, as medical knowledge continues to evolve, it is critical to allow these models to efficiently update outdated or incorrect information without retraining from scratch. Although textual knowledge editing has been widely studied, there is still a lack of systematic benchmarks for multimodal medical knowledge editing involving image and text modalities. To fill this gap, we present MedMKEB, the first comprehensive benchmark designed to evaluate the reliability, generality, locality, portability, and robustness of knowledge editing in medical multimodal large language models. MedMKEB is built on a high-quality medical visual question-answering dataset and enriched with carefully constructed editing tasks, including counterfactual correction, semantic generalization, knowledge transfer, and adversarial robustness. We incorporate human expert validation to ensure the accuracy and reliability of the benchmark. Extensive experiments on state-of-the-art general and medical MLLMs demonstrate the limitations of existing knowledge editing methods in the medical domain, highlighting the need to develop specialized editing strategies.
Zitieren
APA
Xu, D., Wang, J., Chai, Z., Cao, Y., Wang, H., Zhang, H., & Huang, Y. (2026). MedMKEB: A Comprehensive Knowledge Editing Benchmark for Medical Multimodal Large Language Models. Proceedings of the AAAI Conference on Artificial Intelligence, 40(40), 34106-34114. https://doi.org/10.1609/aaai.v40i40.40705
Vancouver
Xu D, Wang J, Chai Z, Cao Y, Wang H, Zhang H, et al. MedMKEB: A Comprehensive Knowledge Editing Benchmark for Medical Multimodal Large Language Models. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(40):34106-34114. doi:10.1609/aaai.v40i40.40705
BibTeX
@article{Xu2026,
author = {Xu, Dexuan and Wang, Jieyi and Chai, Zhongyan and Cao, Yongzhi and Wang, Hanpin and Zhang, Huamin and Huang, Yu},
title = {MedMKEB: A Comprehensive Knowledge Editing Benchmark for Medical Multimodal Large Language Models},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {40},
pages = {34106--34114},
doi = {10.1609/aaai.v40i40.40705}
}
EUVIMED