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

D4

Benchmarks

Task NameDataset NameSOTA ResultTrend
Bivariate Causal DiscoveryD4 s1
Accuracy79
33
ClassificationD4
Mean Accuracy90.062
30
Bivariate Causal DiscoveryD4 s2c
Accuracy64
23
Bivariate Causal DiscoveryD4 s2b
Accuracy63
23
Bivariate Causal DiscoveryD4 s2a
Accuracy71
23
Spatio-temporal signal recoveryD4
MAE1.78
18
Medical Image SegmentationD4
DSC69.04
14
RANKD4 V2
Critical Depth (d50)3
12
Seeker SimulationD4
Precision64.73
12
RegressionD4
RMSE0.2
10
Column Type AnnotationD4-20+
Micro-F187.3
9
Aspect-level sentiment classificationD4
Accuracy85.58
9
RegressionD4
Average Relative MSE0.487
7
Time-Domain PredictionD4
NMSE (dB)-9.3
6
Reliability AssessmentD4 (test)
AU-ARC0.9971
5
Frequency-Domain PredictionD4
NMSE (dB)-16.87
5
Bimanual T-shirt foldingD4 (D1 ∪ DA)
Success Rate100
4
Generative RetrievalD4 (pattern-holdout)
AUC0.6216
4
Generative RetrievalD4 (item-holdout)
AUC0.6561
4
SUCC.D4 V2
Critical Depth (d50)4
4
Compositional ReasoningD4 V2 (test)
Stability (%)89.9
4
CSI ReconstructionD4
NMSE (dB)-18.27
3
Showing 22 of 22 rows