MNIST
Benchmarks
TaskDatasetSOTA resultTrendResults
MNIST
99Average Clustering Accuracy (%)
32
permuted-MNIST (test)
97.61Retained Accuracy
32
split-MNIST (test)
99.66Retained Accuracy
32
MNIST ordered pixel-by-pixel 1.0 (test)
98.8Accuracy
32
MNIST 10K
0.98ARI
31
MNIST (test)
0.932Test Accuracy
31
MNIST (test)
0.003Error Rate
31
MNIST (test)
0.008Test Error
31
Skewed MNIST
95.89Macro Accuracy
30
MNIST
100ARI
30
MNIST
0.056Empirical Epsilon (εα)
30
MNIST
0.16CwECE
30
MNIST
92.6Gauss
30
MNIST
342Compression Ratio (CR)
30
MNIST IID Scenario 1
100F1 Score
30
MNIST 1000 images
79.3Verified Accuracy
30
MNIST (test)
100Coverage
30
MNIST
0.551Kendall Correlation
28
MNIST S2→S0
91.6Accuracy
28
MNIST S0→S2
92.5Accuracy
28
MNIST S1→S0
95.8Accuracy
28
MNIST S0→S1
94.7Accuracy
28
MNIST (test)
0.52Verified Error Rate
28
MNIST
3.82Time Cost (s)
28
pixel-by-pixel MNIST (test)
99.4Accuracy
28