Software by the Wu Lab
Research-grade software for real-world decisions
Deploy robust statistical and AI methods without sacrificing interpretability or scientific validity. These products are designed for high-stakes settings where decisions must be transparent, reproducible, and operationally practical.
Share your use case   |   Explore all repositories   |   Match methods to publications
Choose by goal
Estimate disease burden
Use latent-variable model software for etiologic inference and population-level burden estimation.
Personalize interventions
Use reinforcement-learning and decision-focused tools to support adaptive and individualized care.
Validate AI-supported analysis
Use inference-focused products that remain valid with machine learning predictions or synthetic data.
Loading product highlights...
Why collaborators choose these tools
Maintained and active
Repositories are pulled live from GitHub and prioritized by recent updates to help you find actively maintained options.
Scientifically grounded
Methods are linked to peer-reviewed work and real deployment settings in global and precision health.
Open to collaboration
If your team needs adaptation, benchmarking, or deployment support, reach out and we can scope a joint effort.
Who this is for
Health Researchers
Need valid analysis under missingness, measurement error, or constrained data settings.
Data Scientists
Need methods that are both statistically principled and production-friendly.
Public-Health Teams
Need transparent tools for high-stakes decisions with reproducibility and policy relevance.
Recently Updated Products
Adoption paths
Quick Trial (10-30 mins)
Pick a repo, run the examples, and verify whether assumptions match your setting.
Pilot (1-2 weeks)
Integrate into one workflow, benchmark against your baseline, and document gains.
Deployment Collaboration
For larger efforts, we can discuss tailored onboarding and methodological alignment.
Interested in using a product?
Share your problem setting, data structure, and timeline via zhenkewu [arroba] umich [punto] edu. If useful, include links to your current pipeline or analysis script so we can recommend the best starting point.