Research map / Learning theory and algorithms / Prediction and decision losses
Sparse Concordance-assisted Learning for Optimal Treatment Decision
Shuhan Liang, Wenbin Lu, Rui Song, Lan Wang
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 sparse concordance-assisted learning (SCAL) instead of POWL | treatment rule estimation · Low-dim simulation, n=30 | Estimated value: 1.09 -> 1.16 | Improved |
| use SCAL instead of POWL | treatment rule estimation · Low-dim simulation, n=200 | Estimated value: 1.30 -> 1.32 | No clear effect |
| use SCAL instead of POWL | variable selection · Low-dim simulation, n=200 | MSE of beta: 1.09 -> 0.19 | Improved |
| use SCAL instead of POWL | treatment rule estimation · High-dim randomized, p=500, rho=0, Model 1 | PCD: 0.521 -> 0.744 | Improved |
| use SCAL instead of POWL | treatment rule estimation · High-dim randomized, p=500, rho=0, Model 4 | Estimated value: 3.00 -> 5.67 | Improved |
| use SCAL instead of POWL | treatment rule estimation · High-dim observational, p=500, rho=0.2, Model 1 | PCD: 0.505 -> 0.732 | Improved |
| increase covariate dimension from p=500 to p=1000 | SCAL · High-dim randomized, rho=0, Model 1 | PCD: 0.744 -> 0.741 | Hurt |
| switch interaction from linear (Model 1) to cubic (Model 4) | SCAL · High-dim randomized, p=500, rho=0 | MSE of beta: 0.61 -> 0.35 | Improved |
| switch interaction from linear (Model 1) to cubic (Model 4) | POWL · High-dim randomized, p=500, rho=0 | MSE of beta: 1.81 -> 1.81 | No clear effect |
| use SCAL instead of POWL | treatment rule estimation · STAR*D | Estimated value: -9.46 -> -6.77 | Improved |
| use SCAL instead of always assigning BUP | treatment rule · STAR*D | Estimated value: -10.52 -> -6.77 | Improved |
| use SCAL instead of always assigning SER | treatment rule · STAR*D | Estimated value: -10.74 -> -6.77 | Improved |
Closest papers by meaning
- Learning via Surrogate PAC-Bayes (NeurIPS 2024 · similarity 0.83)
- An Iterative Min-Min Optimization Method for Sparse Bayesian Learning (ICML 2024 · similarity 0.83)
- Incorporating Surrogate Gradient Norm to Improve Offline Optimization Techniques (NeurIPS 2024 · similarity 0.82)
- Trading off Consistency and Dimensionality of Convex Surrogates for Multiclass Classification (NeurIPS 2024 · similarity 0.82)
- Realizable $H$-Consistent and Bayes-Consistent Loss Functions for Learning to Defer (NeurIPS 2024 · similarity 0.82)
- A Universal Growth Rate for Learning with Smooth Surrogate Losses (NeurIPS 2024 · similarity 0.82)
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.
