Mortality prediction on MIMIC-IV (test)
86AUCMedM2T
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MedM2TInput Modality=Vitals + Labs2025.10 | 86 | 45.8 | |
| MedM2TInput Modality=Labs2025.10 | 83 | 39 | |
| IMV-LSTMInput Modality=Vitals + Labs2025.10 | 82.8 | 41.9 | |
| BoXHEDInput Modality=Vitals + Labs2025.10 | 82.6 | 40.6 | |
| MedM2TInput Modality=Vitals2025.10 | 81.8 | 38.7 | |
| IMV-LSTMInput Modality=Labs2025.10 | 81.3 | 39.6 | |
| BoXHEDInput Modality=Labs2025.10 | 79.8 | 36.7 | |
| Bi-LSTMInput Modality=Vitals + Labs2025.10 | 78 | 31 | |
| Bi-LSTMInput Modality=Vitals2025.10 | 76.6 | 31.1 | |
| Bi-LSTMInput Modality=Labs2025.10 | 76.1 | 29.6 | |
| IMV-LSTMInput Modality=Vitals2025.10 | 75.1 | 29.5 | |
| BoXHEDInput Modality=Vitals2025.10 | 74.5 | 28.6 | |
| ReToP_LlamaBackbone=Llama3-8B2026.01 | 71.92 | 6.2 | |
| Flex-MoEMC=L, N, C2024.10 | 69.87 | — | |
| ReToP_QwenBackbone=Qwen2.5-7B2026.01 | 69.75 | 4.61 | |
| MulTMC=L, N, C2024.10 | 67.35 | — | |
| RePrompTsoft prompts (P)=102026.04 | 67.3 | 3.6 | |
| Deepr2026.04 | 66.7 | 2.8 | |
| GRASP2026.04 | 66.6 | 2.9 | |
| StageNet2026.04 | 66.4 | 3.2 | |
| Flex-MoEMC=L, C2024.10 | 66.27 | — | |
| Adore2026.04 | 65.6 | 2.4 | |
| LIMOEMC=L, N, C2024.10 | 65.18 | — | |
| LIMOEMC=L, C2024.10 | 65.14 | — | |
| MulTMC=L, C2024.10 | 65.12 | — | |
| Flex-MoEMC=L, N2024.10 | 64.39 | — | |
| AdaCare2026.04 | 64.3 | 2.9 | |
| Flex-MoEMC=N, C2024.10 | 64.27 | — | |
| BioMedBERTMethod Category=Serialized Classifiers2026.01 | 64.24 | 3.38 | |
| RNNMethod Category=EHR-oriented models2026.01 | 63.33 | 3.38 | |
| GRAM2026.04 | 63.3 | 2.8 | |
| LIMOEMC=L, N2024.10 | 62.64 | — | |
| pi_rfe_dMethod Category=Rewrite-then-Predict models2026.01 | 62.54 | 3.69 | |
| FuseMoEMC=L, C2024.10 | 62.52 | — | |
| MulTMC=N, C2024.10 | 62.18 | — | |
| MAGMC=N, C2024.10 | 61.99 | — | |
| FuseMoEMC=N, C2024.10 | 61.74 | — | |
| FuseMoEMC=L, N, C2024.10 | 61.67 | — | |
| TFMC=N, C2024.10 | 61.37 | — | |
| LIMOEMC=N, C2024.10 | 61.34 | — | |
| MAGMC=L, N, C2024.10 | 61.29 | — | |
| ARCI2026.04 | 61.1 | 3.4 | |
| pi_h_kMethod Category=Rewrite-then-Predict models2026.01 | 61 | 3.58 | |
| MAGMC=L, C2024.10 | 60.75 | — | |
| TFMC=L, N, C2024.10 | 60.66 | — | |
| TFMC=L, C2024.10 | 60.43 | — | |
| RETAIN2026.04 | 60.1 | 3.1 | |
| pi_mrmr_dMethod Category=Rewrite-then-Predict models2026.01 | 60 | 3.21 | |
| LlamaMethod Category=Serialized Classifiers2026.01 | 59.45 | 2.86 | |
| pi_v_oMethod Category=Rewrite-then-Predict models2026.01 | 59.37 | 3.73 | |
| RETAINMethod Category=EHR-oriented models2026.01 | 59.13 | 4.06 | |
| GRASPMethod Category=EHR-oriented models2026.01 | 58.94 | 2.68 | |
| self-genMethod Category=Rewrite-then-Predict models2026.01 | 58.38 | 2.95 | |
| FuseMoEMC=L, N2024.10 | 58.33 | — | |
| MAGMC=L, N2024.10 | 58.11 | — | |
| ModernBERTMethod Category=Serialized Classifiers2026.01 | 58.01 | 3.26 | |
| MulTMC=L, N2024.10 | 57.1 | — | |
| pi_gu_edMethod Category=Rewrite-then-Predict models2026.01 | 56.39 | 2.93 | |
| TFMC=L, N2024.10 | 56.31 | — | |
| pi_b_vMethod Category=Rewrite-then-Predict models2026.01 | 56.05 | 2.57 | |
| pi_f_sMethod Category=Rewrite-then-Predict models2026.01 | 55.82 | 2.63 | |
| QwenMethod Category=Serialized Classifiers2026.01 | 53.87 | 3.12 | |
| REMedMethod Category=EHR-oriented models2026.01 | 52.4 | 2.16 | |
| LlemrMethod Category=EHR-oriented models2026.01 | 52.07 | 3.17 |