Research map / Learning theory and algorithms / Matrix and tensor methods
Distributed High-dimensional Regression Under a Quantile Loss Function
Xi Chen, Weidong Liu, Xiaojun Mao, Zhuoyi Yang
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 |
|---|---|---|---|
| use quantile loss (REL) instead of squared loss Lasso under Cauchy noise | Dist REL · Cauchy noise, n=10000 | l2-error: 128.3 -> 0.168 | Improved |
| use quantile loss (REL) instead of Lasso under Cauchy noise | Dist REL · Cauchy noise, n=10000 | F1-score: 0.27 -> 0.96 | Improved |
| use Lasso instead of REL under normal noise | Lasso · Normal noise, n=10000 | l2-error: 0.102 -> 0.079 | Improved |
| use iterative distributed REL instead of one-shot averaging (Avg-DC) | Dist REL · Cauchy noise, n=10000 | F1-score: 0.11 -> 0.96 | Improved |
| use iterative distributed REL instead of one-shot averaging (Avg-DC) | Dist REL · Normal noise, n=10000 | l2-error: 0.203 -> 0.102 | Improved |
| distribute data over machines instead of pooling (larger first-round regularization) | Dist REL · Normal noise, n=10000 | F1-score: 0.93 -> 0.97 | Mixed |
| increase local sample size m from 200 to 1000 (ablation) | Dist REL · Normal noise, n=5000 | l2-error: 0.154 -> 0.137 | Improved |
| vary bandwidth constant c from 0.5 to 10 (ablation) | Dist REL · Cauchy noise, n=10000 | l2-error: 0.183 -> 0.194 | No clear effect |
| use distributed REL instead of l1-regularized QR solved by LP | Dist REL · Cauchy noise, n=20000 | computation time: 2059.3 -> 0.44 | Improved |
| use distributed REL instead of CSL (smoothed) | Dist REL · Cauchy noise, n=5000, m=200 | l2-error: 0.834 -> 0.251 | Improved |
| increase sparsity level s from 5 to 100 (ablation) | Dist REL · Cauchy noise, n=10000 | l2-error: 0.105 -> 0.330 | Hurt |
Closest papers by meaning
- Distributed High-Dimensional Quantile Regression: Estimation Efficiency and Support Recovery (ICML 2024 · similarity 0.96)
- Communication-Constrained Distributed Quantile Regression with Optimal Statistical Guarantees (JMLR 2022 · similarity 0.91)
- High-Dimensional Structured Quantile Regression (ICML 2017 · similarity 0.88)
- Distributed Learning of Conditional Quantiles in the Reproducing Kernel Hilbert Space (NeurIPS 2022 · similarity 0.87)
- Online Quantile Regression (JMLR 2025 · similarity 0.86)
- Regularized least squares learning with heavy-tailed noise is minimax optimal (NeurIPS 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.
