Research map / Optimization / Implicit bias and generalization
Robust Recovery via Implicit Bias of Discrepant Learning Rates for Double Over-parameterization
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 discrepant learning rates for the matrix and corruption parameters, with α = √n | Double over-parameterization for robust PCA · Robust PCA | relative reconstruction error | Helped |
| Use a single α = √n across varying matrix ranks and corruption ratios | Double over-parameterization for robust PCA · Robust PCA | successful recovery | Helped |
| Use double over-parameterization with discrepant learning rates instead of DIP or DIP-ℓ1 | Robust image recovery · Natural images with salt-and-pepper corruption | PSNR | Helped |
| Continue training without case-dependent early stopping | Double over-parameterization for robust image recovery · Natural images with varying corruption levels | PSNR | Helped |
Closest papers by meaning
- Rank Overspecified Robust Matrix Recovery: Subgradient Method and Exact Recovery (NeurIPS 2021 · similarity 0.90)
- Guarantees of a Preconditioned Subgradient Algorithm for Overparameterized Asymmetric Low-rank Matrix Recovery (ICML 2025 · similarity 0.88)
- Inverse Problems Leveraging Pre-trained Contrastive Representations (NeurIPS 2021 · similarity 0.87)
- Implicit Sparse Regularization: The Impact of Depth and Early Stopping (NeurIPS 2021 · similarity 0.86)
- LERE: Learning-Based Low-Rank Matrix Recovery with Rank Estimation (AAAI 2024 · similarity 0.86)
- Straight-Through Meets Sparse Recovery: the Support Exploration Algorithm (ICML 2024 · similarity 0.85)
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.