active
SDBOR CRG
Neural operators with error estimation and control
Residual-based estimation and correction of neural-operator error in forward simulation, Bayesian inference and topology optimization.
Reliable scientific AI
BayesForSEIRD
Bayesian calibration and model comparison for a continuum SEIRD epidemic model.
Paper companion
Python
Bayesian calibration of a continuum SEIRD epidemic model under uncertainty: parameter inference from reported data, posterior predictive checks, and the model-selection machinery to compare formulations rather than assume one.
Written in Python. It is a paper companion, from the early-pandemic predictive-science work at the Oden Institute, and is included here because the methodology, calibrating a PDE model when the data are noisy and the model is wrong in known ways, is the same one the mechanics work now uses.
Related work
Neural operators and residual correction
Python implementations of neural operators and the residual-correction schemes, with the notebooks used to test them
Publications
- Prashant K. Jha, Lianghao Cao, J. Tinsley Oden, Bayesian-Based Predictions of COVID-19 Evolution in Texas Using Multispecies Mixture-Theoretic Continuum Models, Computational Mechanics, 2020.