Research map / Time series / Causal and generative time series
Dimensionality Reduction for Stationary Time Series via Stochastic Nonconvex Optimization
Minshuo Chen, Lin Yang, Mengdi Wang, Tuo Zhao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Increase downsampling block size from h=1 to h=4 (ablation) | Downsampled Oja with η₀=0.5 · Simulated Gaussian VAR | final principal angle: 0.2320 -> 0.1130 | Improved |
| Increase downsampling block size from h=4 to h=16 (ablation) | Downsampled Oja with η₀=0.5 · Simulated Gaussian VAR | final principal angle: 0.1130 -> 0.3038 | Hurt |
| Use block size h=4 instead of h=1 (ablation) | Downsampled Oja with η₀=0.25 · Simulated Gaussian VAR | final principal angle: 0.3595 -> 0.3745 | No clear effect |
| Use block size h=16 instead of h=4 (ablation) | Downsampled Oja with η₀=1 · Simulated Gaussian VAR | final principal angle: 0.3513 -> 0.5647 | Hurt |
| Use block size h=4 instead of h=1 (ablation) | Downsampled Oja with η₀=0.125 · Simulated Gaussian VAR | final principal angle: 0.7775 -> 0.7892 | No clear effect |
| Use block size h=4 instead of h=1 (ablation) | Downsampled Oja with η₀=2 · Simulated Gaussian VAR | final principal angle: 0.3773 -> 0.4730 | Hurt |
| Set block size to h=4 rather than h=1, 2, 6, 8, or 16 (ablation) | Downsampled Oja with η₀=0.5 · Simulated Gaussian VAR | final principal angle: 0.2320; 0.2080; 0.1130; 0.1287; 0.2828; 0.3038 | Improved |
| Use h=1 rather than h=3 or h=5 (ablation) | Downsampled Oja · Air Quality | projection distortion | Hurt |
| Use h=60 rather than h=3 or h=5 (ablation) | Downsampled Oja · Air Quality | projection distortion | Hurt |
Closest papers by meaning
- Diffusion Approximations for Online Principal Component Estimation and Global Convergence (NeurIPS 2017 · similarity 0.86)
- Streaming PCA for Markovian Data (NeurIPS 2023 · similarity 0.86)
- Latent Diffusion Transformer for Probabilistic Time Series Forecasting (AAAI 2024 · similarity 0.85)
- Non-stationary Diffusion For Probabilistic Time Series Forecasting (ICML 2025 · similarity 0.84)
- Oja's Algorithm for Streaming Sparse PCA (NeurIPS 2024 · similarity 0.84)
- Statistical Guarantees for High-Dimensional Stochastic Gradient Descent (NeurIPS 2025 · 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.
