Research map / Learning theory and algorithms / Prediction and decision losses
More Powerful Conditional Selective Inference for Generalized Lasso by Parametric Programming
Vo Nguyen Le Duy, Ichiro Takeuchi
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 |
|---|---|---|---|
| remove extra conditioning on signs and history by using parametric-programming-based conditional inference instead of over-conditioning | conditional selective inference for generalized lasso · Synthetic data (fused lasso, lasso, elastic net, NNLS, Huber+l1) | TPR (conditional power) and CI length | Improved |
| use parametric programming instead of over-conditioning when accounting for CV selection event | conditional SI for lasso · Synthetic data with lambda from Lambda_1 or Lambda_2 | TPR | Improved |
| use proposed method instead of over-conditioning on real CGH data | fused lasso SI · Dataset 1 resampled from GSE11976 | p-value smaller than OC (% of cases): 55.81% smaller than that of OC | Improved |
| use proposed method instead of over-conditioning on real CGH data | fused lasso SI · Dataset 2 resampled from GSE29172 | p-value smaller than OC (% of cases): 54.24% smaller than that of OC | Improved |
| replace naive sign enumeration with parametric programming | partial-target lasso SI (TN-l1, TN-Custom) · Computational cost comparison | computation time | Improved |
| use proposed method on non-negative least squares versus over-conditioning | NNLS SI · Synthetic NNLS | power | No clear effect |
| use non-normal noise (Laplace, skew normal, t20) or estimated variance | proposed method · Synthetic lasso, 1,200 trials | FPR | No clear effect |
Closest papers by meaning
- More Powerful and General Selective Inference for Stepwise Feature Selection using Homotopy Method (ICML 2021 · similarity 0.89)
- Approximate Post-Selective Inference for Regression with the Group LASSO (JMLR 2023 · similarity 0.88)
- Post-selection inference with HSIC-Lasso (ICML 2021 · similarity 0.88)
- Selective inference for group-sparse linear models (NeurIPS 2016 · similarity 0.87)
- Forward-Backward Selection with Early Dropping (JMLR 2019 · similarity 0.86)
- Kernel-based guarantees for nonlinear parametric models in Bayesian optimization (NeurIPS 2026 · 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.
