Research map / Optimization / Implicit bias and generalization
Global Geometry of Multichannel Sparse Blind Deconvolution on the Sphere
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 manifold gradient descent instead of truncated power iteration for blind gain and phase calibration | MSBD in frequency domain · Synthetic blind calibration, n=128 | recovery phase transition (sparsity s): recovers s = 16 when N >= 80; truncated power iteration fails for s > 8 | Helped |
| use manifold gradient descent instead of truncated power iteration at small N and small s | MSBD in frequency domain · Synthetic blind calibration, n=128 | success rate | Mixed |
| use manifold gradient descent instead of off-the-grid methods on grid-aligned sparse signals | MSBD in frequency domain · Synthetic blind calibration, n=128 | required N for high success rate | Helped |
| increase condition number kappa of f from 1 to 128 (ablation) | manifold gradient descent · Synthetic MSBD, n=256, theta=0.1 | success phase transition in N | No clear effect |
| add Gaussian noise at SNR about 20 dB (ablation) | manifold gradient descent · Synthetic MSBD | number of measurements N for recovery | No clear effect |
| use blind deconvolution instead of non-blind FISTA with a miscalibrated PSF | super-resolution fluorescence microscopy · Microtubule dataset, N=626 | image quality (qualitative) | Helped |
Closest papers by meaning
- On the Global Geometry of Sphere-Constrained Sparse Blind Deconvolution (CVPR 2017 · similarity 0.93)
- A Nonconvex Approach for Exact and Efficient Multichannel Sparse Blind Deconvolution (NeurIPS 2019 · similarity 0.92)
- Structured Local Minima in Sparse Blind Deconvolution (NeurIPS 2018 · similarity 0.91)
- Geometry and Symmetry in Short-and-Sparse Deconvolution (ICML 2019 · similarity 0.88)
- The Price of Sparsity: Sufficient Conditions for Sparse Recovery using Sparse and Sparsified Measurements (NeurIPS 2025 · similarity 0.82)
- Non-Convex Tensor Recovery from Local Measurements (AAAI 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.