Research map / Learning theory and algorithms / Matrix and tensor methods
Estimation in Rotationally Invariant Generalized Linear Models via Approximate Message Passing
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 RI-GAMP with spectral information instead of Gaussian-design GAMP | RI-GAMP · Linear regression and noiseless 1-bit compressed sensing with rotationally invariant design | normalized squared correlation | Helped |
| Use denoisers that depend on all preceding iterates rather than only the latest iterate (ablation) | RI-GAMP · Rotationally invariant design with Beta-distributed eigenvalues | normalized squared correlation | Helped |
| Replace full-history denoisers with latest-iterate denoisers (ablation) | RI-GAMP · Rotationally invariant design with Beta-distributed eigenvalues | normalized squared correlation | Hurt |
| Use RI-GAMP instead of VAMP | RI-GAMP · Linear regression and noiseless 1-bit compressed sensing | normalized squared correlation | Mixed |
| Use RI-GAMP instead of the subgradient method | RI-GAMP · Noiseless 1-bit compressed sensing on a sparse image | normalized squared correlation | Mixed |
| Use RI-GAMP instead of the subgradient method | RI-GAMP · Noiseless 1-bit compressed sensing on a wavelet-transformed RGB image | normalized squared correlation | Mixed |
| Run near the exact-recovery phase transition | RI-GAMP · Noiseless 1-bit compressed sensing | normalized squared correlation | Hurt |
| Increase the number of iterations until covariance matrices become ill-conditioned | RI-GAMP · Rotationally invariant GLMs | numerical stability | Hurt |
Closest papers by meaning
- Inferring Change Points in High-Dimensional Linear Regression via Approximate Message Passing (ICML 2024 · similarity 0.84)
- Matrix Denoising with Doubly Heteroscedastic Noise: Fundamental Limits and Optimal Spectral Methods (NeurIPS 2024 · similarity 0.84)
- Coherence-free Entrywise Estimation of Eigenvectors in Low-rank Signal-plus-noise Matrix Models (NeurIPS 2024 · similarity 0.83)
- Fundamental Limits of Distributed Covariance Matrix Estimation Under Communication Constraints (ICML 2024 · similarity 0.83)
- Nonlinear Laplacians: Tunable principal component analysis under directional prior information (NeurIPS 2025 · similarity 0.83)
- Flat Minima in Linear Estimation and an Extended Gauss Markov Theorem (ICLR 2024 · 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.