Research system

Open-source research systems

Reusable infrastructure for simulation and sequential decision-making. A set of public research systems for virtual Type 1 Diabetes experiments, reinforcement-learning benchmarks and interactive analysis.

Open SourcePyTorchSimulationOffline RLBenchmarking
CAPSML (opens in a new tab)GluCoEnv (opens in a new tab)RL4T1D (opens in a new tab)
GluCoEnv wordmark with a stylized glucose trace

GluCoEnv

GluCoEnv is a PyTorch-based physiological simulation environment designed for reinforcement-learning experiments in glucose control. It supports vectorized virtual subjects, configurable meal scenarios, glucose sensors, insulin pumps, benchmark controllers and reproducible settings.

RL4T1D

RL4T1D organizes agents, environments, clinical comparators, metrics and experiment configuration into a reusable research codebase. It includes infrastructure for both online and offline reinforcement-learning studies.

CAPSML Assistant

CAPSML Assistant makes selected virtual experiments inspectable through the public CAPSML interface. It is a research and education platform for simulation—not a clinical decision-support system—and this site does not display unverified usage statistics.