Embracing Aleatoric Uncertainty in Medical Multimodal Learning with Missing Modalities
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910004939849728 |
|---|---|
| author | Gong, Linxiao Liu, Yang Sun, Lianlong Bi, Yulai Liu, Jing Zhu, Xiaoguang |
| author_facet | Gong, Linxiao Liu, Yang Sun, Lianlong Bi, Yulai Liu, Jing Zhu, Xiaoguang |
| contents | Medical multimodal learning faces significant challenges with missing modalities prevalent in clinical practice. Existing approaches assume equal contribution of modality and random missing patterns, neglecting inherent uncertainty in medical data acquisition. In this regard, we propose the Aleatoric Uncertainty Modeling (AUM) that explicitly quantifies unimodal aleatoric uncertainty to address missing modalities. Specifically, AUM models each unimodal representation as a multivariate Gaussian distribution to capture aleatoric uncertainty and enable principled modality reliability quantification. To adaptively aggregate captured information, we develop a dynamic message-passing mechanism within a bipartite patient-modality graph using uncertainty-aware aggregation mechanism. Through this process, missing modalities are naturally accommodated, while more reliable information from available modalities is dynamically emphasized to guide representation generation. Our AUM framework achieves an improvement of 2.26% AUC-ROC on MIMIC-IV mortality prediction and 2.17% gain on eICU, outperforming existing state-of-the-art approaches. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_21950 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Embracing Aleatoric Uncertainty in Medical Multimodal Learning with Missing Modalities Gong, Linxiao Liu, Yang Sun, Lianlong Bi, Yulai Liu, Jing Zhu, Xiaoguang Machine Learning Medical multimodal learning faces significant challenges with missing modalities prevalent in clinical practice. Existing approaches assume equal contribution of modality and random missing patterns, neglecting inherent uncertainty in medical data acquisition. In this regard, we propose the Aleatoric Uncertainty Modeling (AUM) that explicitly quantifies unimodal aleatoric uncertainty to address missing modalities. Specifically, AUM models each unimodal representation as a multivariate Gaussian distribution to capture aleatoric uncertainty and enable principled modality reliability quantification. To adaptively aggregate captured information, we develop a dynamic message-passing mechanism within a bipartite patient-modality graph using uncertainty-aware aggregation mechanism. Through this process, missing modalities are naturally accommodated, while more reliable information from available modalities is dynamically emphasized to guide representation generation. Our AUM framework achieves an improvement of 2.26% AUC-ROC on MIMIC-IV mortality prediction and 2.17% gain on eICU, outperforming existing state-of-the-art approaches. |
| title | Embracing Aleatoric Uncertainty in Medical Multimodal Learning with Missing Modalities |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2601.21950 |