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

IHDP

Benchmarks

Task NameDataset NameSOTA ResultTrend
Causal InferenceIHDP connected regime
PEHE1.57
72
CATE EstimationIHDP connected regime
PEHE1.57
72
CATE EstimationIHDP 100 train test splits (out-sample)
ERout1.71
53
Individual Treatment Effect EstimationIHDP (within-sample)
Sqrt PEHE0.604
49
Individual Treatment Effect EstimationIHDP (out-of-sample)
sqrt(PEHE)0.613
45
Treatment Effect EstimationIHDP disconnected regime (m1 = 0) (test)
PEHE2.11
40
Treatment Effect EstimationIHDP
PEHE Mean1.392
27
CATE EstimationIHDP (Out-sample)
PEHE0.45
22
CATE EstimationIHDP (In-sample)
PEHE0.41
22
Causal InferenceIHDP
MSE0.0767
20
Potential Outcome EstimationIHDP in-sample
RMSE0.96
19
CATE estimationIHDP Setup D (out-of-sample)
Normalized RMSE0.2
17
CATE estimationIHDP Setup D (in-sample)
Normalized RMSE0.2
17
CATE estimationIHDP Setup C (out-of-sample)
Normalized RMSE0.29
17
CATE estimationIHDP Setup C (in-sample)
NRMSE0.268
17
Conditional distribution estimationIHDP100 a=1 (out-sample)
W2 Score0.021
16
Conditional distribution estimationIHDP100 a=0 (out-sample)
W20.034
16
ATE estimationIHDP
Joint Bias (JB)0.076
14
Causal InferenceIHDP
d_TV (Total Variation Distance)0.239
12
Potential Outcome PredictionIHDP (out-of-sample)
RMSE1.18
11
Causal Effect EstimationIHDP U -> X (test)
Mean Absolute ATE Prop Error53.4
11
Causal Effect EstimationIHDP 40 confounded datasets with no true instrument candidate
Mean Abs ATE Prop Error0.566
9
Causal Effect EstimationIHDP 1000 realizations from NPCI package (heldout)
MAE0.2
9
Causal Effect EstimationIHDP 1000 realizations from NPCI package (train val)
MAE0.14
9
Distributional Fit of Potential OutcomesIHDP (out-of-sample)
Wasserstein Distance0.4
8
Showing 25 of 44 rows