Research map / Bandits and online learning / Online convex optimization and games
Streaming Principal Component Analysis in Noisy Setting
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 |
|---|---|---|---|
| Unbiased randomized gradient estimator for missing entries (MGD-MD) instead of plain MGD on partial data | MGD with missing data · XRMB and MNIST, Bernoulli missingness | variance captured vs number of observations | Helped |
| Oja's algorithm with missing/partial-observation gradient estimators (Oja-MD, Oja-PO) versus full-data Oja (ablation) | Oja's algorithm · XRMB and MNIST with missing data | variance captured vs observations | Hurt |
| Oja-MD/Oja-PO versus GROUSE baseline | Oja with missing entries · XRMB and MNIST | variance captured vs observations | No clear effect |
| Runtime comparison: MGD-PO and MGD-MD versus Oja variants | MGD with missing/partial observations · XRMB and MNIST | progress per runtime | Hurt |
| Using the Oja-style O(dk^2) update instead of MGD | Oja vs MGD (full data) · XRMB and MNIST | progress per runtime | Helped |
| Noisy gradient corruption with bounded total noise (MGD) | MGD · online regret theory | regret bound | No clear effect |
| Robust symmetrized absolute-deviation loss (abs-MGD) vs squared loss | MGD with outliers · theory, adversarial sequences | regret bound | No clear effect |
Closest papers by meaning
- Robust Streaming PCA (NeurIPS 2022 · similarity 0.93)
- Nearly-Linear Time and Streaming Algorithms for Outlier-Robust PCA (ICML 2023 · similarity 0.92)
- Low Precision Streaming PCA (NeurIPS 2025 · similarity 0.89)
- Oja's Algorithm for Streaming Sparse PCA (NeurIPS 2024 · similarity 0.86)
- Capturing the denoising effect of PCA via compression ratio (NeurIPS 2024 · similarity 0.85)
- Temporal Streaming Batch Principal Component Analysis for Time Series Classification (Student Abstract) (AAAI 2025 · similarity 0.84)
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.