Research map / Learning theory and algorithms / Sample complexity
A Fast and Accurate Estimator for Large Scale Linear Model via Data Averaging
Rui Wang, Yanyan Ouyang, Panpan Yu, Wangli Xu
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 |
|---|---|---|---|
| replace random grouping (VDA) with selected/ordered grouping of observations (ablation) | data averaging estimator · Case 2, p=200 | empirical MSE (x10^3): 1034.62 -> 69.4913 | Improved |
| replace random grouping (VDA) with selected grouping (ablation) | data averaging estimator · Case 1, p=200 | empirical MSE (x10^3): 15014 -> 2469.56 | Improved |
| replace random grouping (VDA) with selected grouping (ablation) | data averaging estimator · Case 4, p=200 | empirical MSE (x10^3): 345.624 -> 10.658 | Improved |
| data averaging into 2p groups instead of IBOSS sampling | linear model sketching · Case 3, p=200 | empirical MSE (x10^3): 235.444 -> 23.0075 | Improved |
| data averaging into 2p groups instead of IBOSS sampling | linear model sketching · Case 4, p=200 | empirical MSE (x10^3): 37.7542 -> 10.658 | Improved |
| data averaging into 2p groups instead of IBOSS sampling | linear model sketching · Case 1, p=50 | empirical MSE (x10^3): 507.642 -> 182.438 | Improved |
| data averaging into 2p groups instead of IBOSS sampling | linear model sketching · Case 4, p=50 | empirical MSE (x10^3): 2.36205 -> 2.00812 | Improved |
| data averaging instead of uniform sampling | linear model sketching · Case 2, p=200 | empirical MSE (x10^3): 1033.37 -> 69.4913 | Improved |
| increase dimension p from 50 to 200 for the proposed method | NEW · Case 1 | empirical MSE (x10^3): 182.438 -> 2469.56 | Hurt |
| use data averaging instead of leverage-score sampling | linear model sketching · Case 5, p=200 | empirical MSE (x10^3): 1979.68 -> 192.085 | Improved |
| use data averaging instead of IBOSS for computing time | NEW · N=8×10^4, p=200 time | computing time (s): 0.16388 -> 0.21261 | No clear effect |
| use data averaging instead of uniform sampling for computing time | NEW · N=8×10^4, p=200 time | computing time (s): 0.00653 -> 0.21261 | Hurt |
Closest papers by meaning
- A Statistical Perspective on Randomized Sketching for Ordinary Least-Squares (JMLR 2016 · similarity 0.89)
- Distributed Least Squares in Small Space via Sketching and Bias Reduction (NeurIPS 2024 · similarity 0.87)
- Sketched Ridgeless Linear Regression: The Role of Downsampling (ICML 2023 · similarity 0.87)
- Sketched Ridge Regression: Optimization Perspective, Statistical Perspective, and Model Averaging (ICML 2017 · similarity 0.85)
- Matrix Product Sketching via Coordinated Sampling (ICLR 2025 · similarity 0.85)
- Sketching for Convex and Nonconvex Regularized Least Squares with Sharp Guarantees (ICLR 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.
