End-to-end reconstruction meets data-driven regularization for inverse problems

Our work derives ideas from learned optimization and adversarial machine learning, and makes an attempt to combine the best features of both aforementioned paradigms.