Weakly-Supervised Multimodal Learning on MIMIC-CXR

Andrea Agostini, Daphné Chopard, Yang Meng, Norbert Fortin, Babak Shahbaba +3 more
11/15/2024
cs.LG

No official code found. Best implementation: agostini335/mmvmvae-mimic (6 stars, GPL-3.0, last commit 2024-12-05).

Abstract

Multimodal data integration and label scarcity pose significant challenges for machine learning in medical settings. To address these issues, we conduct an in-depth evaluation of the newly proposed Multimodal Variational Mixture-of-Experts (MMVM) VAE on the challenging MIMIC-CXR dataset. Our analysis demonstrates that the MMVM VAE consistently outperforms other multimodal VAEs and fully supervised approaches, highlighting its strong potential for real-world medical applications.

View on arXivView PDF

Code Implementations(1)

agostini335/mmvmvae-mimicCommunity100%
GPL-3.0
60Python, Jupyter NotebookOct 29, 20241 years agoGPL-3.0

Similar Papers

PiCME: Pipeline for Contrastive Modality Evaluation and Encoding in the MIMIC Dataset

82% match2025

Unified Cross-Modal Medical Image Synthesis with Hierarchical Mixture of Product-of-Experts

81% match1 repo162024

Aggregation of Dependent Expert Distributions in Multimodal Variational Autoencoders

81% match1 repo42025

Med-MoE: Mixture of Domain-Specific Experts for Lightweight Medical Vision-Language Models

80% match1 repo1582024

Towards Robust Multimodal Representation: A Unified Approach with Adaptive Experts and Alignment

80% match1 repo62025

I2MoE: Interpretable Multimodal Interaction-aware Mixture-of-Experts

80% match1 repo742025

Cite this paper

@article{agostini2024weaklysupervised,
  title  = {Weakly-Supervised Multimodal Learning on MIMIC-CXR},
  author = {Andrea Agostini and Daphné Chopard and Yang Meng and Norbert Fortin and Babak Shahbaba and Stephan Mandt and Thomas M. Sutter and Julia E. Vogt},
  year   = {2024},
  eprint = {2411.10356},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2411.10356v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2411.10356.svg)](https://codeofpaper.com/paper/2411.10356)

Discussion