A recent study in Nature Medicine demonstrates how diffusion models can enhance the robustness and fairness of medical machine learning models in interpreting medical images like histopathology, chest X-rays, and dermatological images by augmenting training datasets with synthetic data, improving accuracy and fairness across diverse patient groups.
Study investigates the impact of AI on clinicians' diagnostic accuracy, revealing that AI explanations don't significantly mitigate the harmful effects of biased AI models, emphasizing the need for careful validation of AI tools in clinical settings.
A study by the University of the Basque Country (UPV/EHU) uses game theory to establish that tumors with less cellular heterogeneity are more aggressive.