Search Results for author:
Found 4 papers, 2 papers with code
Saliency is a Possible Red Herring When Diagnosing Poor Generalization
In some prediction tasks, such as for medical images, one may have some images with masks drawn by a human expert, indicating a region of the image containing relevant information to make the prediction.
Underwhelming Generalization Improvements From Controlling Feature Attribution
We describe a simple method for taking advantage of such auxiliary labels, by training networks to ignore the distracting features which may be extracted outside of the region of interest, on the training images for which such masks are available.
Deep neural network or dermatologist?
We show that despite high accuracy, the models will occasionally assign importance to features that are not relevant to the diagnostic task.
GradMask: Reduce Overfitting by Regularizing Saliency
With too few samples or too many model parameters, overfitting can inhibit the ability to generalise predictions to new data.
