Research map / Reinforcement learning / Policy gradient and actor-critic
Experimental design for MRI by greedy policy search
Tim Bakker, Herke van Hoof, Max Welling
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use greedy one-step reward objective instead of non-greedy (gamma 1.0) (ablation) | policy-gradient MRI sampler · Knee long horizon | SSIM: 0.7421 -> 0.7442 | Improved |
| use greedy objective instead of non-greedy (ablation) | policy-gradient MRI sampler · Brain long horizon | SSIM: 0.8886 -> 0.8917 | Improved |
| use greedy objective instead of non-greedy (ablation) | policy-gradient MRI sampler · Knee base horizon | SSIM: 0.7223 -> 0.7230 | Improved |
| discount rewards with gamma 0.9 instead of 1.0 (ablation) | non-greedy policy · Knee long horizon | SSIM: 0.7421 -> 0.7449 | Improved |
| gamma 0.9 versus greedy (ablation) | policy-gradient MRI sampler · Knee base horizon | SSIM: 0.7230 -> 0.7232 | No clear effect |
| adaptive greedy policy versus non-adaptive oracle | Greedy policy · Knee long horizon | SSIM: 0.7421 -> 0.7442 | Improved |
| non-greedy policy versus non-adaptive oracle | NGreedy policy · Brain long horizon | SSIM: 0.8909 -> 0.8886 | Hurt |
| greedy objective instead of non-greedy | policy gradient · Knee long horizon, epoch 20 | gradient SNR: 1.04 -> 3.49 | Improved |
| longer horizon for non-greedy estimator | NGreedy · Knee epoch 20 | gradient SNR: 1.24 -> 1.04 | Hurt |
| gamma 0.9 instead of 1.0 on long horizon (ablation) | non-greedy estimator · Brain long epoch 20 | gradient SNR: 1.18 -> 7.22 | Improved |
| AlphaZero search policy versus NGreedy | AlphaZero baseline · Knee base horizon | SSIM: 0.7223 -> 0.7203 | Hurt |
| larger or smaller policy architecture (ablation) | Greedy and NGreedy policies · Knee | SSIM | No clear effect |
Closest papers by meaning
- Experimental Design for Multi-Channel Imaging via Task-Driven Feature Selection (ICLR 2024 · similarity 0.84)
- Progressive Divide-and-Conquer via Subsampling Decomposition for Accelerated MRI (CVPR 2024 · similarity 0.84)
- Hierarchical Variational Policies for Reward-Guided Diffusion (NeurIPS 2026 · similarity 0.84)
- Optimizing ADMM and Over-Relaxed ADMM Parameters for Linear Quadratic Problems (AAAI 2024 · similarity 0.84)
- A Unified Model for Compressed Sensing MRI Across Undersampling Patterns (CVPR 2025 · similarity 0.84)
- Efficient Noise Calculation in Deep Learning-based MRI Reconstructions (ICML 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
