Twins
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Twins RCT
1.372Multiplicative Calibration Error
15
Twins
0.092Qini Coefficient
15
TWINS
0Mean Effect
15
Twins (in-sample)
0.87AUC
13
Twins (out-sample)
86.1AUC
13
Twins Posterior
0Mean BSE
10
Twins censoring Real (test)
2.63RMSE (L=30)
9
Twins no censoring Real (test)
2.53RMSE (L=30)
9
TWINS HeavyTails outcome (synthetic)
0.25Average W1 Error
6
Twins
0.031RMSE
6
Twins (n=3200)
0.061MAE (eATE)
6
Twins (n=1600)
0.061Mean Absolute ATE Error (eATE)
6
Twins n=200
0.065MAE (eATE)
6
Twins (n=50)
0.062Mean Absolute ATE Error (eATE)
6
Twins Prior
0.039Mean BSE
5
Twins (out-sample)
0.861AUC
5
Twins (in-sample)
0.87AUC
5
Twins RealCause (test)
99.8Empirical Coverage
3
Twins
68.748Runtime (s/it)
2
Twins
97.12Coverage Mean
1