Self-Supervised Aggregation of Diverse Experts for Test-Agnostic Long-Tailed Recognition

We theoretically show that our self-supervised strategy has a provable ability to simulate test-agnostic class distributions.