Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition from a Domain Adaptation Perspective
This work addresses the challenge of improving model performance on tail classes in long-tailed datasets, which is crucial for real-world applications with imbalanced data distributions, though it is incremental as it builds on existing class-balanced methods.
The paper tackles the problem of long-tailed visual recognition by analyzing class-balanced methods from a domain adaptation perspective, revealing their limitations in handling tail classes, and proposes a meta-learning approach to estimate distribution differences, achieving improved performance validated on six benchmark datasets.
Object frequency in the real world often follows a power law, leading to a mismatch between datasets with long-tailed class distributions seen by a machine learning model and our expectation of the model to perform well on all classes. We analyze this mismatch from a domain adaptation point of view. First of all, we connect existing class-balanced methods for long-tailed classification to target shift, a well-studied scenario in domain adaptation. The connection reveals that these methods implicitly assume that the training data and test data share the same class-conditioned distribution, which does not hold in general and especially for the tail classes. While a head class could contain abundant and diverse training examples that well represent the expected data at inference time, the tail classes are often short of representative training data. To this end, we propose to augment the classic class-balanced learning by explicitly estimating the differences between the class-conditioned distributions with a meta-learning approach. We validate our approach with six benchmark datasets and three loss functions.