EHR-MPC: Inference-Time Control for Sepsis Treatment with Generative Patient Digital Twins

2026-07-13T07:23:56Z41c462ee82c5e3a67b88cd07ce843053ef030b9451990b3f0c0a3c3f89374a12
EHRadversarial-mlbanditscausal-inferenceclinical-decision-supportdeconfoundingdeep-gaussian-processesdigital-twinhealthcarehigh-dimensional-statisticsmachine-learningmedical-aimodel-predictive-controlonline-learningpatient-safetyprivacyreinforcement-learningrobustnesssepsis

What happened

Collection of newly announced ML papers (arXiv 2026-07-13) focused on medical decision-making, high-dimensional inference, robustness to confounding, structured probabilistic models, and bandit/online learning. Notable item: EHR-MPC introduces a generative EHR "digital twin" plus model-predictive control to plan sepsis treatments at inference time and is evaluated on a multicenter ICU cohort (Mass General Brigham). Security-relevant concerns include patient-safety and clinical risk from automated treatment planning, privacy and data-governance risks from use of real EHR data, model robustness/

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
arxiv_stat_ml
Record identifier
41c462ee82c5e3a67b88cd07ce843053ef030b9451990b3f0c0a3c3f89374a12
Enrichment time
2026-07-13T07:23:56Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.

Record · EHR-MPC: Inference-Time Control for Sepsis Treatment with Generative Patient Digital Twins · Baitaphish