We investigated the relations between the Shapley values and the labels. We asked three radiologists (R.Y., S.R. and D.L.R.) to re-label 300 chest X-rays (i.e. 100 least valuable, 100 most valuable and 100 randomly sampled from the training set). The radiologists were blinded to the original labels and the Shapley values while labeling the chest X-rays. For each chest X-ray, the radiologists had the choice of labeling it as pneumonia, no pneumonia or unsure.

In order to further evaluate mislabeled images and their Shapley values, we asked the three radiologists to re-evaluate the 100 most valuable and the 100 least valuable chest X-ray images, and determined whether these images have other types of abnormalities (e.g. opacity) or are completely normal.

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.