Theoretical Perspectives on Deep Learning Methods in Inverse Problems

In recent years, there have been significant advances in the use of deep learning methods inverse problems such as denoising, compressive sensing, inpainting, and super-resolution. While this line works has predominantly driven by practical algorithms experiments, it also given rise to a variety intriguing theoretical problems. paper, we survey some prominent developments works, focusing particular on generative priors, untrained neural network unfolding algorithms. addition summarizing existing results these topics, highlight several ongoing challenges open