Live intelligence

Latest AI updates

2 updates
Research

ICare-DT: A Transformer Digital Twin for Hourly ICU Forecasting

Researchers propose ICare-DT, a Transformer-based digital twin that updates a latent patient state hourly from new intensive care measurements without retraining. Trained on 140,473 eICU-CRD stays, later forecasts improved over the twin's own hour-6 forecasts in all 12 settings, with gains of +0.073 to +0.153 AUROC; validation controls linked the improvement to patients' own new measurements. Without refitting, the model retained gains on a second database in 8 of 9 settings (+0.025 to +0.114 AUROC) in a prespecified MIMIC-IV evaluation. The work also introduces a benchmark for treatment-response prediction, finding that standard learners did not outperform predicting zero effect for an out-of-support dosing regime, while an informed learner recovered simulated effects with R2 0.30 to 0.40 at one effect size.

Digital twinsIntensive careClinical forecasting
Research

EpiCoach: Auditable LLM Agent for N-of-1 Wearable Physiology

EpiCoach is an LLM-agent system that treats a single cross-country mountain biker as the subject of auditable N-of-1 science, keeping claims in an evidence-gated hypothesis registry that requires a probabilistic prediction about a future session before it occurs. Forecasts come from a mechanistic digital twin built on heart rate, GPS and barometric altitude, combining a lagged heart-rate response model, a fitness-fatigue model with learned time constants and a latent technical-descending skill state measured on discovered trail segments. The agent prescribes each training week as a constrained experiment that maximizes information gain subject to load and safety constraints, and is evaluated over eighteen months of data with a rolling-origin replay benchmark and a live prospective phase. In the first three live weeks the registry used 20 predictions and scored 14, the twin beat standard forecasters on heart rate but not segment time, and the safety officer vetoed the agent twice; the registry, benchmark and protocol are released.

LLM agents for healthdigital twinswearable sensing