Research map / Optimization / Optimizers
Optimization-Derived Learning with Essential Convergence Analysis of Training and Hyper-training
Risheng Liu, Xuan Liu, Shangzhi Zeng, Jin Zhang, Yixuan Zhang
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 |
|---|---|---|---|
| Jointly optimize training and hyper-training variables with BMO | Sparse coding · Set14 | PSNR: 15.59±0.81 -> 18.82±1.59; 15.64±0.87 -> 18.98±2.53 | Improved |
| Jointly optimize training and hyper-training variables with BMO | Sparse coding · Set14 | SSIM: 0.52±0.13 -> 0.63±0.16; 0.52±0.13 -> 0.65±0.15 | Improved |
| Compose the numerical PG operator with DRUNet in BMO | Image deconvolution · Set3c, six image/noise settings | PSNR (dB): Butterfly, σ=1%: 34.18 -> 33.67; Leaves, σ=1%: 35.12 -> 35.39; Starfish, σ=1%: 33.91 -> 33.98; Butterfly, σ=3%: 29.45 -> | Mixed |
| Use BMO with the RCDNet iterative architecture | Rain streak removal · Rain100L | PSNR: 40.00 -> 40.07 | Improved |
| Use BMO with the RCDNet iterative architecture | Rain streak removal · Rain100H | PSNR: 31.28 -> 30.96 | Hurt |
| Use BMO with the RCDNet iterative architecture | Rain streak removal · Rain100H | SSIM: 0.903 -> 0.905 | Improved |
| Enforce a non-expansive learned mapping (ablation) | BMO sparse coding · Sparse coding | gradient of upper loss φ: null | Improved |
Closest papers by meaning
- Learning Provably Improves the Convergence of Gradient Descent (NeurIPS 2025 · similarity 0.86)
- A Stochastic Approach to Bi-Level Optimization for Hyperparameter Optimization and Meta Learning (AAAI 2025 · similarity 0.85)
- Unleashing the Power of Meta-tuning for Few-shot Generalization Through Sparse Interpolated Experts (ICML 2024 · similarity 0.84)
- Memory-Reduced Meta-Learning with Guaranteed Convergence (AAAI 2025 · similarity 0.84)
- qNBO: quasi-Newton Meets Bilevel Optimization (ICLR 2025 · similarity 0.84)
- Scalable Meta-Learning via Mixed-Mode Differentiation (ICML 2025 · similarity 0.84)
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.
