GE

Benchmarks

Task Name	Dataset Name	SOTA Result
Classification	GE	Accuracy66.5	37
Classification	GE 50% corrupted features (50% noise)	Accuracy (50% Noise)60	21
Downstream performance prediction	GE German corrupted with Laplace noise	Downstream Performance0.6	19
Feature Selection	GE with random extraneous features	Precision100	17
Feature Selection	GE with random extraneous features	ROC-AUC100	17
Tabular Prediction	GE	Score0.7	17
Classification	GE 75% corrupted features noise	Accuracy57.9	10
Tabular Classification	GE	Macro F163.8	6
Synthetic Tabular Data Generation	GE	DCR23.4	2

Showing 9 of 9 rows