Tabular Synthetic Data Generation on Adult low-order real-world (test)
84.41AccuracyUB
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| UBepsilon=1.0, Description=Upper bound performance trained on real data2026.06 | 84.41 | 75.68 | 0.014 | 0.027 | 0.041 | 94.5 | 94.09 | |
| AIMepsilon=1.02026.06 | 83.36 | 76.1 | 0.017 | 0.034 | 0.053 | 87.06 | 93.18 | |
| PrivMRFepsilon=1.02026.06 | 83.15 | 76.85 | 0.017 | 0.033 | 0.052 | 84.15 | 93.74 | |
| Tab-PEepsilon=1.02026.06 | 82.22 | 73.66 | 0.049 | 0.086 | 0.118 | 34.27 | 77.25 | |
| PrivGSDepsilon=1.02026.06 | 82.09 | 75.9 | 0.017 | 0.033 | 0.052 | 74.45 | 80.81 | |
| RAP++epsilon=1.02026.06 | 80.87 | 72.22 | 0.023 | 0.043 | 0.066 | 61.08 | 80.64 | |
| GEMepsilon=1.02026.06 | 79.17 | 69.63 | 0.037 | 0.073 | 0.109 | 0.185 | 76.48 | |
| PrivSynepsilon=1.02026.06 | 75.77 | 43.11 | 0.011 | 0.034 | 0.061 | 45.34 | 89.34 |