No abstract

In this work, we train conditional Wasserstein generative adversarial networks to effectively sample from the posterior of physics-based Bayesian inference problems. The generator is constructed using a U-Net architecture, with the latent information injected using conditional instance normalization. The former facilitates a multiscale inverse map, while the latter enables the decoupling of the latent space dimension from the dimension of the measurement, and introduces stochasticity at all scales of the U-Net. We solve PDE-based inverse problems to demonstrate the performance of our approach in quantifying the uncertainty in the inferred field. Further, we show the generator can learn inverse maps which are local in nature, which in turn promotes generalizability when testing with out-of-distribution samples.

No abstract

Summary We present an efficient adjoint‐based framework for computing sensitivities of quantities of interest with respect to material parameters for coupled fluid‐structural acoustic systems with explicit interface coupling. The fluid is modeled using the Helmholtz equation and the structure is modeled using the Navier‐Cauchy equations. Sensitivities are used to drive a gradient based optimization algorithm to solve important problems in structural acoustics, viz noise minimization and vibration isolation. For each problem, we consider two different priors: one where the optimal solution has a smooth variation and another with a bimaterial distribution. These priors are imposed with the help of suitable regularization terms. The effectiveness of this approach is demonstrated on both interior and exterior structural acoustic problems.