Stress-Testing of Multimodal Models in Medical Image-Based Report Generation

Abstract

Multimodal models, namely vision-language models, present unique possibilities through the seamless integration of different information mediums for data generation. These models mostly act as a black-box, making them lack transparency and explicability. Reliable results require accountable and trustworthy Artificial Intelligence (AI), namely when in use for critical tasks, such as the automatic generation of medical imaging reports for healthcare diagnosis. By exploring stress-testing techniques, multimodal generative models can become more transparent by disclosing their shortcomings, further supporting their responsible usage in the medical field.

Cite

Text

Carvalhido et al. "Stress-Testing of Multimodal Models in Medical Image-Based Report Generation." AAAI Conference on Artificial Intelligence, 2025. doi:10.1609/AAAI.V39I28.35203

Markdown

[Carvalhido et al. "Stress-Testing of Multimodal Models in Medical Image-Based Report Generation." AAAI Conference on Artificial Intelligence, 2025.](https://mlanthology.org/aaai/2025/carvalhido2025aaai-stress/) doi:10.1609/AAAI.V39I28.35203

BibTeX

@inproceedings{carvalhido2025aaai-stress,
  title     = {{Stress-Testing of Multimodal Models in Medical Image-Based Report Generation}},
  author    = {Carvalhido, Flávia and Cardoso, Henrique Lopes and Cerqueira, Vítor},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2025},
  pages     = {29251-29252},
  doi       = {10.1609/AAAI.V39I28.35203},
  url       = {https://mlanthology.org/aaai/2025/carvalhido2025aaai-stress/}
}