For a long time, there has been a strong set of criticisms of deep learning methods in the community due to the lack of interpretability, insight, reliability, and the fact that they are data and computational resource hungry. The new framework of deep unfolding promises a solution to many of such issues in deep learning by relying on the unfolding of the iterations of well-established model-based optimization or inference algorithms onto the layers of deep neural networks—presenting a unique gateway to model-based deep learning. Given their recent emergence, the study of deep unfolding networks (DUNs) is somewhat limited to developing their architectures in various signal processing and computing applications. To unleash its true potential, my research aims to establish the theoretical foundations, unveil the game-changing opportunities, and address several fundamental questions associated with model-based deep learning.