AlphaNet: Improving Long-Tail Classification By Combining Classifiers
This work addresses the challenge of improving classification for rare classes in long-tailed datasets, which is crucial for real-world applications where data distribution is imbalanced, though it is incremental as it builds on existing models.
The paper tackles the problem of misclassifying rare classes as visually similar frequent classes in long-tail learning, introducing AlphaNet, a post hoc correction method that updates rare class classifiers using linear combinations of frequent class classifiers, which greatly improves test accuracy for rare classes with minimal impact on overall accuracy.
Methods in long-tail learning focus on improving performance for data-poor (rare) classes; however, performance for such classes remains much lower than performance for more data-rich (frequent) classes. Analyzing the predictions of long-tail methods for rare classes reveals that a large number of errors are due to misclassification of rare items as visually similar frequent classes. To address this problem, we introduce AlphaNet, a method that can be applied to existing models, performing post hoc correction on classifiers of rare classes. Starting with a pre-trained model, we find frequent classes that are closest to rare classes in the model's representation space and learn weights to update rare class classifiers with a linear combination of frequent class classifiers. AlphaNet, applied to several models, greatly improves test accuracy for rare classes in multiple long-tailed datasets, with very little change to overall accuracy. Our method also provides a way to control the trade-off between rare class and overall accuracy, making it practical for long-tail classification in the wild.