Deep long-tailed learning, one of the most challenging problems in visual recognition, aims
to train well-performing deep models from a large number of images that follow a long-tailed …

Standard training via empirical risk minimization (ERM) can produce models that achieve
low error on average but high error on minority groups, especially in the presence of …

Abstract Machine learning algorithms typically assume that training and test examples are
drawn from the same distribution. However, distribution shift is a common problem in real …

Distribution shifts—where the training distribution differs from the test distribution—can
substantially degrade the accuracy of machine learning (ML) systems deployed in the wild …

Self-supervised learning (SSL) is a scalable way to learn general visual representations
since it learns without labels. However, large-scale unlabeled datasets in the wild often have …

A fundamental limitation of applying semi-supervised learning in real-world settings is the
assumption that unlabeled test data contains only classes previously encountered in the …

Modern deep neural networks (DNNs) become frail when the datasets contain noisy
(incorrect) class labels. Robust techniques in the presence of noisy labels can be …

While deep learning models have shown remarkable performance in various tasks, they are
susceptible to learning non-generalizable _spurious features_ rather than the core features …

Modern neural networks have the capacity to overfit noisy labels frequently found in real-
world datasets. Although great progress has been made, existing techniques are very …

Abstract Self-supervised Contrastive Learning (CL) has been recently shown to be very
effective in preventing deep networks from overfitting noisy labels. Despite its empirical …