Enhancing high-dimensional probabilistic model updating: A generic generative model-inspired framework with GAN-embedded implementation

Jiang Mo, Wang-Ji Yan, Ka-Veng Yuen, Michael Beer

Abstract: Highlights•An innovative PMU framework that emulates the core principles of generative model is proposed.•Model parameters of the GMM input sampler are embedded in an interpretable network as learnable parameters by reparameterization trick.•A learnable MMD discriminator as the distance metric is devised to achieve a more nuanced measurement of distribution disparity.•Adversarial training enhances the generator’s generative power and the discriminator’s discernment capability.•High-dimensional PMU is performed in an efficient network training manner with stochastic gradient descent.

External IDs:doi:10.1016/j.cma.2025.118190