Neural collapse provides an elegant mathematical characterization of learned last layer
representations (aka features) and classifier weights in deep classification models. Such …

Abstract Open-world Semi-Supervised Learning aims to classify unlabeled samples utilizing
information from labeled data while unlabeled samples are not only from the labeled known …

Incremental learning models need to update the categories and their conceptual
understanding over time. The current research has placed more emphasis on learning new …

Neural collapse ($\mathcal {NC} $) is a phenomenon observed in classification tasks where
top-layer representations collapse into their class means, which become equinorm …

Recent studies empirically demonstrate the positive relationship between the transferability
of neural networks and the in-class variation of the last layer features. The recently …

Deep neural networks typically require a large number of accurately labeled images for
training with cross-entropy loss, and often overfit noisy labels. Contrastive learning has …

The current paradigm of training deep neural networks for classification tasks includes
minimizing the empirical risk that pushes the training loss value towards zero, even after the …

Deep learning has achieved impressive performance across various medical imaging tasks.
However, its inherent bias against specific groups hinders its clinical applicability in …

Label noise is a common issue in real-world datasets that inevitably impacts the
generalization of models. This study focuses on robust classification tasks where the label …

Link prediction on graphs has achieved great success with the rise of deep graph learning.
However, the potential robustness under the edge noise is less investigated. We reveal that …