# Experimental design for MRI by greedy policy search

arXiv: https://arxiv.org/abs/2010.16262 · HTML: https://mlsanity.com/paper/2010.16262/

## Key facts

- Cited by 40 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Oct 2020.

## Built on by (most cited first, 30 of 40)

- [A comprehensive survey on deep active learning in medical image analysis](https://mlsanity.com/paper/2310.14230/) (2023)
- [Active 3D Shape Reconstruction from Vision and Touch](https://mlsanity.com/paper/2107.09584/) (2021)
- [On learning adaptive acquisition policies for undersampled multi-coil MRI reconstruction](https://mlsanity.com/paper/2203.16392/) (2022)
- [End-to-End Sequential Sampling and Reconstruction for MRI](https://mlsanity.com/paper/2105.06460/) (2021)
- [Adaptive Compressed Sensing with Diffusion-Based Posterior Sampling](https://mlsanity.com/paper/2407.08256/) (2024)
- [On Retrospective k-space Subsampling schemes For Deep MRI Reconstruction](https://mlsanity.com/paper/2301.08365/) (2023)
- [Learning Optimal K-space Acquisition and Reconstruction using Physics-Informed Neural Networks](https://mlsanity.com/paper/2204.02480/) (2022)
- [E-Valuating Classifier Two-Sample Tests](https://mlsanity.com/paper/2210.13027/) (2022)
- [Active Diffusion Subsampling](https://mlsanity.com/paper/2406.14388/) (2024)
- [Optimal MRI Undersampling Patterns for Ultimate Benefit of Medical Vision Tasks](https://mlsanity.com/paper/2108.04914/) (2021)
- [Adaptive Sampling of k-Space in Magnetic Resonance for Rapid Pathology Prediction](https://mlsanity.com/paper/2406.04318/) (2024)
- [Sequential Posterior Sampling with Diffusion Models](https://mlsanity.com/paper/2409.05399/) (2024)
- [Towards Ultrafast MRI via Extreme k-Space Undersampling and Superresolution](https://mlsanity.com/paper/2103.02940/) (2021)
- [Towards Robust Active Feature Acquisition](https://mlsanity.com/paper/2107.04163/) (2021)
- [Accelerated Intravascular Ultrasound Imaging using Deep Reinforcement Learning](https://mlsanity.com/paper/2201.09522/) (2022)
- [Diffusion Active Learning: Towards Data-Driven Experimental Design in Computed Tomography](https://mlsanity.com/paper/2504.03491/) (2025)
- [JoJoNet: Joint-contrast and Joint-sampling-and-reconstruction Network for Multi-contrast MRI](https://mlsanity.com/paper/2210.12548/) (2022)
- [On the Feasibility of Machine Learning Augmented Magnetic Resonance for Point-of-Care Identification of Disease](https://mlsanity.com/paper/2301.11962/) (2023)
- [Indirectly Parameterized Concrete Autoencoders](https://mlsanity.com/paper/2403.00563/) (2024)
- [End-to-end Adaptive Dynamic Subsampling and Reconstruction for Cardiac MRI](https://mlsanity.com/paper/2403.10346/) (2024)
- [The MRI Scanner as a Diagnostic: Image-less Active Sampling](https://mlsanity.com/paper/2406.16754/) (2024)
- [Adaptive Selection of Sampling-Reconstruction in Fourier Compressed Sensing](https://mlsanity.com/paper/2409.11738/) (2024)
- [Active Phase-Encode Selection for Slice-Specific Fast MR Scanning Using a Transformer-Based Deep Reinforcement Learning Framework](https://mlsanity.com/paper/2203.05756/) (2022)
- [A Path Towards Clinical Adaptation of Accelerated MRI](https://mlsanity.com/paper/2208.12835/) (2022)
- [On Sensitivity and Robustness of Normalization Schemes to Input Distribution Shifts in Automatic MR Image Diagnosis](https://mlsanity.com/paper/2306.13276/) (2023)
- [Learning to sample in Cartesian MRI](https://mlsanity.com/paper/2312.04327/) (2023)
- [Variational Learning ISTA](https://mlsanity.com/paper/2407.06646/) (2024)
- [Active Sampling for MRI-based Sequential Decision Making](https://mlsanity.com/paper/2505.04586/) (2025)
- [Online Feedback Efficient Active Target Discovery in Partially Observable Environments](https://mlsanity.com/paper/2505.06535/) (2025)
- [Replacing thinking with tool usage enables reasoning in small language models](https://mlsanity.com/paper/2507.05065/) (2025)

## Abstract (arXiv)

In today's clinical practice, magnetic resonance imaging (MRI) is routinely accelerated through subsampling of the associated Fourier domain. Currently, the construction of these subsampling strategies - known as experimental design - relies primarily on heuristics. We propose to learn experimental design strategies for accelerated MRI with policy gradient methods. Unexpectedly, our experiments show that a simple greedy approximation of the objective leads to solutions nearly on-par with the more general non-greedy approach. We offer a partial explanation for this phenomenon rooted in greater variance in the non-greedy objective's gradient estimates, and experimentally verify that this variance hampers non-greedy models in adapting their policies to individual MR images. We empirically show that this adaptivity is key to improving subsampling designs.