De-identified Medical Data
A federated-learning–based, privacy-preserving computation platform for standardized cleaning and structured storage of multi-center medical data — data that is usable but never exposed.
Rebuilding carbon-based life in silicon. We build trustworthy medical AI and, through simulation of virtual mitochondria and digital organs, explore the algorithmic nature of life’s intelligence.
CORE RESEARCH MATRIX & INFRASTRUCTURE
A federated-learning–based, privacy-preserving computation platform for standardized cleaning and structured storage of multi-center medical data — data that is usable but never exposed.
Energy-metabolism modeling of artificial mitochondria and signaling-pathway simulation of virtual cells. We build high-fidelity digital-twin organoids for drug screening and pathology inference.
We release domain foundation models for medicine and maintain OpenMedicalGym, a simulation environment that gives developers worldwide a “digital proving ground” for testing algorithm robustness.
From perception to cognition. We develop medical agents capable of autonomous decision-making, enabling human–AI collaboration in assisted diagnosis, surgical planning and digital-organ maintenance.
INSTITUTE MILESTONES
Launching simulation projects for digital heart and liver organoids with tens of millions of parameters, exploring a new paradigm for non-invasive clinical trials.
Won First Prize at the NHC Digital Health Innovation Application Competition and runner-up at the AWS Hackathon, establishing early impact in the field.
Established the IMPF-AI vision and completed the first round of de-identified data collection and the core computing platform.
HONORS & COMPETITIONS
2nd National Digital Health Innovation Application Competition
Best Demo Award
AWS Healthcare AI Competition
We are looking for partners and researchers passionate about computational biology, medical AI and complex-systems simulation. Medical institutions are also welcome to join our de-identified data network.