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

real-world datasets

Benchmarks

Task NameDataset NameSOTA ResultTrend
ClassificationReal-world datasets Aggregate Mean
G-Mean88.5
54
Prototype SelectionReal-world datasets aggregated (test)
Average Rank1.93
54
Node Classification9 real-world datasets (average)
Average Ranking1.56
43
Local Feature Importance Evaluation12 Real-World Datasets Aggregate (test)
Average Rank1
16
Tabular Imputation20 Real-world Datasets Overall 2026 (test)
Mean NRMSE0.163
12
Tabular Imputation20 real-world datasets MCAR 2026 (test)
NRMSE (5% Missing Data)0.118
12
Tabular Imputation20 real-world datasets MNAR 2026 (test)
NRMSE (5% Missing Rate)0.146
12
Feature Importance ExplanationReal-world datasets
Best Score97.8
12
AI-generated image detectionReal-world Datasets Chameleon, SynthWildX, WildRF Aggregate
Accuracy95.8
11
Reflective surface reconstructionReal-world datasets
Time (h)0.5
10
Video Motion MagnificationReal-world Datasets
MANIQA Score (Baby Scene)0.7475
10
Single Image Reflection RemovalFive Real-world Datasets (Real20, Objects, Postcard, Wild, Nature) Average (494) (test)
PSNR27.21
9
Prototype Selection Stability15 real-world datasets 10-fold stratified CV (test)
Mean Jaccard22.5
8
Low-light image enhancementReal-world datasets (MEF, LIME, DICM, NPE) (test)
MEF Score4.14
6
Showing 14 of 14 rows