Image Classification on Fashion-MNIST class-prior shift ρ=50, μ=0.5 (test)
78.09AccuracyDIW
Evaluation results
| Method | Links | |
|---|---|---|
| DIWTraining Time (s)=22.97, Imbalance ratio (ρ)=50, Fraction of minority classes (μ)=0.5, Backbone=LeNet-5, Optimizer=SGD2026.05 | 78.09 | |
| ADIW-KLTraining Time (s)=8.39, Imbalance ratio (ρ)=50, Fraction of minority classes (μ)=0.5, Backbone=LeNet-5, Optimizer=SGD2026.05 | 77.86 | |
| ADIW-KMTraining Time (s)=6.84, Imbalance ratio (ρ)=50, Fraction of minority classes (μ)=0.5, Backbone=LeNet-5, Optimizer=SGD2026.05 | 77.69 | |
| ADIW-WATraining Time (s)=7.64, Imbalance ratio (ρ)=50, Fraction of minority classes (μ)=0.5, Backbone=LeNet-5, Optimizer=SGD2026.05 | 77.37 | |
| ADIW-LSTraining Time (s)=7.67, Imbalance ratio (ρ)=50, Fraction of minority classes (μ)=0.5, Backbone=LeNet-5, Optimizer=SGD2026.05 | 76.93 | |
| Val-OnlyTraining Time (s)=0.02, Imbalance ratio (ρ)=50, Fraction of minority classes (μ)=0.5, Backbone=LeNet-5, Optimizer=SGD2026.05 | 65.55 |