Research map / Learning theory and algorithms / Matrix and tensor methods
Operational Support Estimator Networks
Mete Ahishali, Mehmet Yamaç, Serkan Kiranyaz B, Moncef Gabbouj
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 |
|---|---|---|---|
| replace convolutional neurons with operational super neurons (OSEN1 vs CSEN1) | CSEN1 · MNIST support recovery MR=5% | F1-score (%): 78.86 -> 86.77 | Improved |
| replace CSEN2 by OSEN2 | CSEN2 · MNIST support recovery MR=5% | F1-score (%): 83.37 -> 88.13 | Improved |
| add non-causal learning (NCL) scheme to OSEN1 (ablation) | OSEN1 · MNIST support recovery MR=5% | F1-score (%): 86.77 -> 90.37 | Improved |
| add NCL to OSEN2 (ablation) | OSEN2 · MNIST support recovery MR=5% | F1-score (%): 88.13 -> 90.39 | Improved |
| increase polynomial order Q from 1 to 5 (ablation) | OSEN1 · MNIST support recovery MR=5% | F1-score (%): 84.38 -> 87.34 | Improved |
| increase Q from 1 to 5 (ablation) | NCL-OSEN1 · MNIST support recovery MR=25% | F1-score (%): 95.77 -> 95.81 | No clear effect |
| double hidden neurons in CSEN (CSEN2+ vs CSEN2) (ablation) | CSEN2 · MNIST support recovery MR=5% | F1-score (%): 83.37 -> 85.82 | Improved |
| proposed OSEN versus CSEN at low measurement rate | OSEN · Yale-B face recognition MR=0.01 | accuracy: accuracy gap around 25% | Improved |
| add transposed operational layers (second configuration) (ablation) | OSEN · Yale-B face recognition | accuracy | No clear effect |
| use OSEN prior with weighted TV instead of plain TV | CS MRI reconstruction · MRI MR=0.25 | PSNR (dB): 43.23 -> 45.20 | Improved |
| use OSEN prior instead of CSEN prior | CS MRI reconstruction · MRI MR=0.05 | PSNR (dB): 30.05 -> 30.49 | Improved |
| OSEN1 versus CSEN1 in MRI gradient-domain support recovery | OSEN1 · MRI MR=5% | F1-score (%): 52.13 -> 53.94 | Improved |
Closest papers by meaning
- Straight-Through Meets Sparse Recovery: the Support Exploration Algorithm (ICML 2024 · similarity 0.85)
- Deep Compressed Sensing (ICML 2019 · similarity 0.84)
- Training Image Estimators without Image Ground Truth (NeurIPS 2019 · similarity 0.84)
- Iterative Regularization with k-support Norm: An Important Complement to Sparse Recovery (AAAI 2024 · similarity 0.83)
- Untrained Neural Nets for Snapshot Compressive Imaging: Theory and Algorithms (NeurIPS 2024 · similarity 0.82)
- NPN: Non-Linear Projections of the Null-Space for Imaging Inverse Problems (NeurIPS 2025 · similarity 0.82)
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.
