| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| ICD Coding | MIMIC-III 50 labels (test) | F1 Micro64.1 | 70 | |
| Counterfactual Outcome Prediction | MIMIC-III semi-synthetic (N=3000) (test) | RMSE0.24 | 35 | |
| Counterfactual Outcome Prediction | MIMIC-III semi-synthetic (N=1000) (test) | RMSE0.3 | 35 | |
| Relation Extraction | MIMIC-III (test) | Strict F186.7 | 26 | |
| Named Entity Recognition | MIMIC-III (test) | Strict F185.6 | 26 | |
| ICD coding | MIMIC-III full (test) | F1 Micro59.9 | 19 | |
| Irregular Time Series Forecasting | MIMIC-III (h) 36 → 12 horizon | MSE1.44 | 18 | |
| Irregular Time Series Forecasting | MIMIC-III 24 → 24 horizon | MSE1.63 | 18 | |
| Irregular Time Series Forecasting | MIMIC-III 12 → 36 horizon | MSE1.8 | 18 | |
| Clinical time-series prediction | MIMIC-III (test) | AUROC85.79 | 18 | |
| ICD-9 code prediction | MIMIC-III 8922 labels (full) | AUC Macro0.91 | 17 | |
| Disease Diagnosis | MIMIC-III 1.0 (test) | Micro Precision69.04 | 17 | |
| De-identification | MIMIC-III 100 discharge notes 1.4 (test) | Precision99 | 14 | |
| Medical Code Prediction | MIMIC-III full-label 1.4 (test) | F1 Micro0.586 | 14 | |
| Missing data estimation | MIMIC-III v1.4 (test) | Mean RMSE0.0141 | 13 | |
| Overall Diagnosis Prediction | MIMIC-III 5% (train) | Precision@10 (Visit-Level)51.73 | 11 | |
| Medication Recommendation | MIMIC-III (test) | Jaccard Similarity55.77 | 10 | |
| Phenotype Classification | MIMIC-III v1.4 (test) | F1 Score40.13 | 10 | |
| Length of Stay | MIMIC-III v1.4 (test) | F1 Score66.39 | 10 | |
| 48-hour In-Hospital Mortality | MIMIC-III v1.4 (test) | F1 Score53.44 | 10 | |
| Length of Stay | MIMIC-III | AUROC68.2 | 10 | |
| Length of Stay | MIMIC-III 72-hour observation horizon | AUROC0.661 | 10 | |
| Mortality Prediction | MIMIC-III target v1.4 | AUPRC16.95 | 10 | |
| ICD coding | MIMIC-III Full v1.4 (test) | Macro F10.123 | 10 | |
| In-hospital mortality prediction | MIMIC-III Hypertension | AUPRC47.27 | 10 |