Integrating data distribution prior via Langevin dynamics for end‐to‐end MR reconstruction

To develop a novel deep learning-based method inheriting the advantages of data distribution prior and end-to-end training for accelerating MRI.