Advancing diagnostic performance and clinical usability of neural networks via adversarial training and dual batch normalization
Sign Up to like & getrecommendations! Published in 2021 at "Nature Communications"
DOI: 10.1038/s41467-021-24464-3
Abstract: Unmasking the decision making process of machine learning models is essential for implementing diagnostic support systems in clinical practice. Here, we demonstrate that adversarially trained models can significantly enhance the usability of pathology detection as… read more here.
Keywords: diagnostic performance; batch normalization; batch; advancing diagnostic ... See more keywords