Research map / Learning theory and algorithms / Matrix and tensor methods
High-Dimensional L2-Boosting: Rate of Convergence
Ye Luo, Martin Spindler, Jannis Kueck
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 |
|---|---|---|---|
| add post-selection OLS refit (ablation) | L2Boosting with data-dependent stopping · Riboflavin production | out-of-sample MSE: 0.3641 -> 0.1237 | Improved |
| add orthogonal projection after each step (ablation) | L2Boosting with data-dependent stopping · Riboflavin production | out-of-sample MSE: 0.3641 -> 0.1080 | Improved |
| use proposed data-dependent stopping instead of Ks rule (ablation) | L2Boosting · Riboflavin production | out-of-sample MSE: 0.4669 -> 0.3641 | Improved |
| add post-selection refit (ablation) | LASSO · Riboflavin production | out-of-sample MSE: 0.1687 -> 0.1539 | Improved |
| add post-selection refit (ablation) | L2Boosting · Student scores (Mathematics) | out-of-sample MSE: 19.1 -> 19.3 | No clear effect |
| add post-selection refit (ablation) | L2Boosting · Student scores (Portuguese) | out-of-sample MSE: 8.0 -> 7.9 | No clear effect |
| add post-selection refit (ablation) | L2Boosting (data-dependent stopping) · Simulation, sparse iid (n=200, p=200) | out-of-sample MSE: 0.35 -> 0.21 | Improved |
| add orthogonal projection (ablation) | L2Boosting (data-dependent stopping) · Simulation, sparse iid (n=200, p=200) | out-of-sample MSE: 0.35 -> 0.24 | Improved |
| use data-dependent stopping instead of Ks rule (ablation) | L2Boosting · Simulation, sparse iid (n=200, p=200) | out-of-sample MSE: 0.41 -> 0.35 | Improved |
| add post-selection refit (ablation) | LASSO with cross-validation · Simulation, sparse iid (n=200, p=200) | out-of-sample MSE: 0.30 -> 0.44 | Hurt |
| add post-selection refit (ablation) | LASSO · Simulation, polynomial iid (n=200, p=200) | out-of-sample MSE: 0.34 -> 0.52 | Hurt |
| add post-selection refit (ablation) | L2Boosting (data-dependent stopping) · Simulation, polynomial iid (n=200, p=200) | out-of-sample MSE: 0.39 -> 0.42 | Hurt |
Closest papers by meaning
- Early stopping for kernel boosting algorithms: A general analysis with localized complexities (NeurIPS 2017 · similarity 0.86)
- SnapBoost: A Heterogeneous Boosting Machine (NeurIPS 2020 · similarity 0.86)
- A Practical Scheme and Fast Algorithm to Tune the Lasso With Optimality Guarantees (JMLR 2016 · similarity 0.86)
- Statistical Inference for Gradient Boosting Regression (NeurIPS 2025 · similarity 0.86)
- Post-selection inference with HSIC-Lasso (ICML 2021 · similarity 0.86)
- The High Line: Exact Risk and Learning Rate Curves of Stochastic Adaptive Learning Rate Algorithms (NeurIPS 2024 · similarity 0.85)
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.
