Criteo
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| CTR Prediction | Criteo | AUC0.8163 | 309 | |
| CTR Prediction | Criteo (test) | AUC0.8158 | 147 | |
| Click-Through Rate Prediction | Criteo (val) | Validation Loss0.4515 | 60 | |
| Click-Through Rate Prediction | Criteo (test) | AUC0.8114 | 57 | |
| Dimension Selection | Criteo | AUC80.14 | 49 | |
| Click-Through Rate Prediction | Criteo | AUC77.197 | 44 | |
| Click-Through Rate Prediction | Criteo standard (test) | AUC0.812 | 24 | |
| Conditional Average Treatment Effect Estimation | Criteo RCT | Multiplicative Calibration Error1.044 | 15 | |
| Uplift Modeling | Criteo | Qini Coefficient85.2 | 15 | |
| Binary Classification | Criteo | Total Time (s)59.49 | 14 | |
| Delayed Conversion Rate Prediction | Criteo | AUC0.843 | 11 | |
| Click-through rate prediction | Criteo x1 (test) | AUC0.815 | 11 | |
| CVR prediction | Criteo (streaming) | AUC100 | 7 | |
| Click-Through Rate (CTR) Prediction | Criteo WSDM 2021 (test) | LogLoss0.4403 | 7 | |
| CTR Prediction | Criteo (offline) | AUC0.7991 | 7 | |
| Ad Ranking | Criteo (test) | AUC0.7842 | 6 | |
| Uplift Modeling | Criteo stratified 50k (subsample) | Normalized Qini Score100 | 5 | |
| VFL Attack Performance | Criteo | LISR67.55 | 4 | |
| Click-through-rate prediction | Criteo-1B | Test AUC0.6269 | 4 | |
| Optimization Convergence | Criteo (train) | Final Training Loss Gradient Magnitude0.033 | 3 | |
| Binary Classification | Criteo (test) | AUC0.8038 | 3 | |
| CTR Prediction | Criteo Ad Kaggle (test) | Accuracy78.76 | 2 | |
| Conditional Average Treatment Effect (CATE) Estimation | Semi-Synthetic Criteo N=61,200 (subsampled) | Improvement98.6 | 1 |
Showing 23 of 23 rows