A physics informed surrogate model for elliptic PDEs and its Bayesian inverse problem analysis

Ponente(s): Antonio Capella Kort

The talk discusses Bayesian inference in inverse problems with uncertainty quantification, where solving partial differential equations is computationally expensive. To reduce this cost, a new physics-informed surrogate model is proposed for linear elliptic PDEs. The study shows the model's consistency and effectiveness through numerical examples with synthetic data, significantly speeding up computation from months to minutes with minimal loss of accuracy.