Inverse Modeling

We address inverse problems, which may lack unique solutions but can be transformed into well-posed problems through deep learning-based regularization and the incorporation of physics constraints. There, analyzing and tuning the latent space representation is one of the used approaches.