Differentiable Bayesian Inversion

A modular computational framework that couples geological priors, physical or learned simulators and measurement-error models inside one differentiable workflow, so Bayesian inference over subsurface models can be driven by automatic differentiation.