Deep learning for inverse problems with unknown operator

Miguel del Alamo

We consider ill-posed inverse problems where the forward operator $T$ is unknown, and instead we have access to training data consisting of functions $f_i$ and their noisy images $…