Abstract:
Learning low-dimensional features that are both diverse and discriminative from high-dimensional data is a fundamental and critical task. This paper proposes a novel loss...Show MoreMetadata
Abstract:
Learning low-dimensional features that are both diverse and discriminative from high-dimensional data is a fundamental and critical task. This paper proposes a novel loss/principle for feature learning inspired by the "oil and water separation" phenomenon. The features learned with the proposed loss are shown to be more robust to label corruptions in empirical evaluations. Moreover, the proposed method achieves state-of-the-art performance in certain downstream tasks.
Date of Conference: 08-11 October 2023
Date Added to IEEE Xplore: 11 September 2023
ISBN Information: