Label-Imbalanced and Group-Sensitive Classification under Overparameterization
–Neural Information Processing Systems
Classical methods, such as weighted cross-entropy, fail when training deep nets to the terminal phase of training (TPT), that is training beyond zero training error.
Neural Information Processing Systems
Feb-10-2026, 06:42:56 GMT
- Country:
- North America
- Canada > British Columbia (0.04)
- United States > California
- Riverside County > Riverside (0.04)
- Santa Barbara County > Santa Barbara (0.04)
- North America
- Genre:
- Research Report > New Finding (0.68)
- Technology: