Deep Long-Tailed Learning: A Survey

TLDR

Deep long‑tailed learning seeks to train high‑performing deep models from data with a long‑tailed class distribution, yet class imbalance often biases models toward dominant classes and hampers tail‑class performance, prompting extensive research over the past decade that has yielded significant progress. This survey aims to comprehensively review recent advances in deep long‑tailed learning. The authors categorize existing methods into class re‑balancing, information augmentation, and module improvement, review each in detail, and empirically evaluate their effectiveness using a newly proposed relative‑accuracy metric. The survey concludes by outlining key applications of deep long‑tailed learning and proposing several promising research directions.

Abstract

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 class distribution. In the last decade, deep learning has emerged as a powerful recognition model for learning high-quality image representations and has led to remarkable breakthroughs in generic visual recognition. However, long-tailed class imbalance, a common problem in practical visual recognition tasks, often limits the practicality of deep network based recognition models in real-world applications, since they can be easily biased towards dominant classes and perform poorly on tail classes. To address this problem, a large number of studies have been conducted in recent years, making promising progress in the field of deep long-tailed learning. Considering the rapid evolution of this field, this paper aims to provide a comprehensive survey on recent advances in deep long-tailed learning. To be specific, we group existing deep long-tailed learning studies into three main categories (i.e., class re-balancing, information augmentation and module improvement), and review these methods following this taxonomy in detail. Afterward, we empirically analyze several state-of-the-art methods by evaluating to what extent they address the issue of class imbalance via a newly proposed evaluation metric, i.e., relative accuracy. We conclude the survey by highlighting important applications of deep long-tailed learning and identifying several promising directions for future research.