Decision support and principled inference for partially observed systems across earth, environmental, and health sciences — determining what actions to take, what experiments to run, and what measurements are worth collecting when interventions are costly and uncertainty is unavoidable.
Website · Publications · Projects
- Decision support under partial observability — surveillance design, forecasting pipelines, intervention evaluation, resource allocation
- Model criticism & evaluation — structured observables, Pareto-optimal configuration selection, Bayesian stacking, proper scoring rules
- Scientific AI/ML — physics-embedded surrogates, lawful learning, generative model design
- Operator-partitioned solvers — IMEX/PDE operator splitting, trait-structured dynamical systems
- Bayesian latent variable models — variational PCA, posterior predictive testing, rank selection
- Operational forecasting — infectious disease scenario modeling, ensemble calibration, intervention timing
| Package | Language | Description |
|---|---|---|
| model-criticism | Python | Observable-based model evaluation, Pareto optimization, Bayesian stacking |
| ModelCriticism.jl | Julia | Model evaluation harness wrapping Metaheuristics.jl, ParetoSmooth.jl, QuasiMonteCarlo.jl, GlobalSensitivity.jl |
| Package | Org | Description |
|---|---|---|
| VBPCApy | yoavram-lab | Variational Bayesian PCA with missing-data support |
| pp-eigentest | yoavram-lab | Posterior predictive eigenvalue rank testing |
| op_engine | ACCIDDA | Operator-partitioned solver (Python) |
| op_system | ACCIDDA | System specification compiler (Python) |
| flepimop2 | ACCIDDA | Modular epidemic modeling and simulation pipeline |