Multiscale Bayesian framework bridging 5-species oral biofilm ecology and 3D FEM stress analysis via TMCMC, DeepONet surrogate (~100x speedup), and GNN-informed priors. 28x stiffness range, basin sensitivity, gradient-based NUTS sampling.
bayesian-inference uncertainty-quantification finite-element-method jax tmcmc graph-neural-network inverse-problem multiscale-modeling oral-biofilm neural-operator deeponet peri-implantitis nuts-sampler hamilton-principle dysbiosis-index
-
Updated
Mar 9, 2026 - Python