[论文解读] Deep Long-Tailed Learning: A Survey

Yifan Zhang, Bingyi Kang+3 more|arXiv (Cornell University)|Oct 9, 2021

Domain Adaptation and Few-Shot Learning参考文献 166被引用 79

一句话总结

对深度长尾学习的全面综述，将方法分为类别再平衡、信息增强和模块改进，并提出一种用于评估的新相对准确率度量。

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.

研究动机与目标

• 总结在长尾类别分布下训练深度模型的挑战。
• 将现有的深度长尾学习方法分为三个主要类别和九个子类别。
• 提供一个基于新的评估指标（relative accuracy）的实证分析框架，用于评估方法如何应对数据不平衡。
• 突出实际应用并确定未来研究的有前景方向。

提出的方法

• 将并评估现有方法整理为三个主要类别：类再平衡、信息增强和模块改进。
• 在类再平衡方面，讨论重新采样、对类别敏感的学习和对数i t 调整。
• 评述信息增强，包括迁移学习和数据增强。
• 涵盖模块改进方法，如表征学习、分类器设计、解耦训练和集成学习。
• 提出并应用一种新的评估指标 relative accuracy，用于分析最先进方法。

实验结果

研究问题

• RQ1 现有的深度长尾学习方法如何在头部、中部和尾部类别之间缓解不平衡分布？
• RQ2 在三个广义类别（再平衡、信息增强、模块改进）中，哪些核心技术能为尾部类别带来最佳提升？
• RQ3 使用所提议的 relative accuracy 指标评估时，当前方法的效果如何？
• RQ4 哪些数据集和应用能体现这些方法的实际影响与局限性？
• RQ5 进一步推进深度长尾学习最具前景的未来方向和新的任务设定有哪些？

主要发现

• 提供到2021年中为止对深度长尾学习的首个全面综述。
• 对先进方法的深入评述以及使用 relative accuracy 指标的实证分析，以评估不平衡处理。
• 指出四个潜在的方法创新方向以及未来研究的八种新任务设定。
• 讨论在细粒度、不平衡视觉任务中的实际应用及现有方法的局限性。