| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Mortality Prediction | eICU | AUC-ROC0.9373 | 22 | |
| Diagnosis Prediction | eICU spatial domain shift, Midwest as target (test) | Weighted F169.72 | 21 | |
| Urinary tract infection (UTI) prediction | eICU | AUROC0.945 | 19 | |
| Sepsis Prediction | eICU | AUROC98.2 | 19 | |
| EHR prediction | eICU | AUROC0.78 | 16 | |
| Heart Failure Prediction | eICU spatial domain shift 2014-2015 (test) | AUROC92.54 | 15 | |
| Medication Recommendation | eICU | Jaccard Similarity41.95 | 13 | |
| Vasopressor (Vaso) use prediction | eICU | AUROC0.77 | 12 | |
| Length of Stay (LOS) prediction | eICU | AUROC70 | 12 | |
| Clinical utility evaluation | eICU | Micro-averaged AUROC87 | 10 | |
| Readmission Prediction | eICU spatial domain shift, Midwest as target (test) | AUPRC22.74 | 10 | |
| Safety Compliance Evaluation | eICU-AC | LPA98.4 | 10 | |
| Shock Prediction | eICU | Mean Balanced Accuracy65.7 | 9 | |
| Readmission Prediction | eICU | AUC-ROC0.7879 | 9 | |
| Survival Analysis (Sepsis) | EICU | Ctd0.87 | 8 | |
| Data Imputation | eICU MNAR 50% missingness | NRMSE0.148 | 8 | |
| Data Imputation | eICU MNAR 30% missingness | NRMSE0.145 | 8 | |
| Data Imputation | eICU MNAR 10% missingness | NRMSE0.125 | 8 | |
| Prolonged Length of Stay (PLOS) Prediction | eICU | F1 Score72.49 | 7 | |
| Length of Stay Prediction | eICU | MAD1.78 | 7 | |
| Imputation | eICU MAR 30% missingness | NRMSE0.125 | 7 | |
| Imputation | eICU MCAR 30% missingness | NRMSE0.124 | 7 | |
| ICU mortality prediction | eICU YAIB-aligned | AUROC86.4 | 7 | |
| Uplift Modeling | eicu | AUROC0.54 | 6 | |
| Synthetic Data Generation | eICU (test) | MMD0.012 | 6 |