Stylist: Style-Driven Feature Ranking for Robust Novelty Detection
The article presents STYLIST, a new method for novelty detection, which is the identification of meaningful deviations from established data distributions. It separates semantic changes, which are relevant to the…
Continue reading