Research map / Learning theory and algorithms / Matrix and tensor methods
Coherence Pursuit: Fast, Simple, and Robust Subspace Recovery
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 single-step Coherence Pursuit (one Gram matrix multiplication with p-norm ranking) instead of iterative robust PCA solvers | robust PCA subspace recovery · Running time, m=n=10000 | run time (s): 27 s vs 401 s (FMS), 3547 s (OP), 598 s (R1-PCA) | Helped |
| Adopt the l1-norm (p=1) ranking for CoP when outliers are repeated (ablation) | CoP with linearly dependent outliers · Repeated outliers illustration (r=5, n1=50, n2=500) | separation of inlier vs outlier scores in p | Mixed |
| Use l2-norm (p=2) instead of l1 for ranking coherence (ablation) | CoP · Theoretical sufficient conditions (Lemmas 1-2) | outlier tolerance | Helped |
| Adaptive column sampling (Algorithm 2) for subspace identification with random projection | CoP · Subspace identification step | redundant column sampling | Helped |
| Use CoP as error-correction step on subspace clustering output (Hopkins155 traffic) | subspace clustering error correction · Hopkins155 traffic sequences, 20% initial clustering error | clustering error | Helped |
Closest papers by meaning
- Dual Principal Component Pursuit for Robust Subspace Learning: Theory and Algorithms for a Holistic Approach (ICML 2021 · similarity 0.89)
- Accelerated Combinatorial Search for Outlier Detection with Provable Bound on Sub-Optimality (AAAI 2021 · similarity 0.89)
- Fast and Provable Algorithms for Sparse PCA with Improved Sample Complexity (ICML 2025 · similarity 0.89)
- Capturing the denoising effect of PCA via compression ratio (NeurIPS 2024 · similarity 0.88)
- Efficient Sparse PCA via Block-Diagonalization (ICLR 2025 · similarity 0.87)
- Oja's Algorithm for Streaming Sparse PCA (NeurIPS 2024 · similarity 0.87)
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.