Research
Preserving explainability in federated learning for clinical decision-making: SHAP explainability consistency analysis of ICU mortality prediction models
Nature ·
The Brief
- A nature.com article examines explainability consistency in federated learning for clinical decision-making.
- The study applies SHAP explainability consistency analysis to ICU mortality prediction models.
- The source provided no summary, so findings beyond the title are not reported here.
Why it matters
Clinicians need explanations of model predictions to trust them, and this work examines whether those explanations stay consistent when models are trained across federated data sources.
Who’s affected
- clinicians and researchers working on federated learning for ICU care
- developers of explainable AI tools for clinical decision support
- #federated learning
- #explainable ai
- #shap
- #icu mortality
- #clinical decision support