Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

F-MNIST

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image ClassificationF-MNIST (test)
Accuracy97.62
173
Image ClassificationF-MNIST
Accuracy92.8
149
Backdoor AttackF-MNIST
ASR97.81
25
Conformal PredictionF-MNIST
Average Prediction Set Size1.7
24
INR classificationF-MNIST Implicit Neural Representations (test)
Accuracy84.6
21
Image ClassificationF-MNIST 30% Minority (test)
Majority Accuracy94.34
20
Image ClassificationF-MNIST 10% Minority (test)
Majority Accuracy90.32
20
Differentially Private Image SynthesisF-MNIST
FID11.6
16
Image ClassificationF-MNIST NIID, κ=1
Accuracy74.89
12
Image ClassificationF-MNIST Uniform
Accuracy84.45
12
Image ClassificationF-MNIST IDN-50% (test)
Accuracy76.43
12
Image ClassificationF-MNIST IDN-40% (test)
Accuracy85.81
12
Image ClassificationF-MNIST IDN-30% (test)
Accuracy89.61
12
Image ClassificationF-MNIST IDN-20% (test)
Accuracy90.83
12
Image ClassificationF-MNIST IDN-10% (test)
Accuracy91.96
12
One-class classificationf-MNIST (test)
AUC95.8
12
ClassificationF.MNIST last 5% of data stream
Accuracy (F.MNIST last 5%)75.22
10
Stream ClassificationF.MNIST last 20% of data stream
Average Accuracy74.85
10
ClassificationF.MNIST last 10% data stream
Average Accuracy74.73
10
Dimensionality ReductionF-MNIST
Triplet Centroid Accuracy92.5
10
Dimensionality ReductionF-MNIST (test)
Class Angular Distortion Index0.0207
8
Dimensionality ReductionF-MNIST
Distance Correlation0.957
7
Semantic Communication Backdoor AttackF-MNIST standard (test)
Attack Success Rate (ASR)99.98
7
Classification under DBA attackF-MNIST
Balance Accuracy96.74
7
Image ClassificationF-MNIST 1000x1
Accuracy83
6
Showing 25 of 49 rows