Research map / Learning theory and algorithms / Prediction and decision losses
Doubly Debiased Robust Subsampling for Transfer Learning
Tao Wang, Weng Kee Wong
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 target refinement to IPW subsampling (ablation) | Doubly debiased robust subsampling · DGP1 covariate shift, Delta=0.4, nS=50K nT=2K | test MSE: 1.2107 -> 1.1429 | Improved |
| add IPW debiasing to target refinement (ablation) | Doubly debiased robust subsampling · DGP1 covariate shift, Delta=0.4, nS=50K nT=2K | test MSE: 1.1764 -> 1.1429 | Improved |
| add target refinement to IPW subsampling (ablation) | Doubly debiased robust subsampling · DGP2 label shift, severe, nS=50K nT=2K | misclassification error (%): 25.1862 -> 22.8763 | Improved |
| add IPW debiasing to target refinement (ablation) | Doubly debiased robust subsampling · DGP2 label shift, severe, nS=50K nT=2K | misclassification error (%): 24.6237 -> 22.8763 | Improved |
| add target refinement to IPW subsampling (ablation) | Doubly debiased robust subsampling · DGP3 concept shift, zeta=0.2, nS=50K nT=2K | test MSE: 1.3296 -> 1.1964 | Improved |
| add IPW debiasing to target refinement (ablation) | Doubly debiased robust subsampling · DGP3 concept shift, zeta=0.2, nS=50K nT=2K | test MSE: 1.2586 -> 1.1964 | Improved |
| add target refinement to IPW subsampling (ablation) | Doubly debiased robust subsampling · DGP4 heavy tails, nu=2, nS=50K nT=2K | test MSE: 2.0326 -> 1.7826 | Improved |
| use the proposed method instead of uniform subsampling | Doubly debiased robust subsampling · DGP1 covariate shift, Delta=0.4, nS=50K nT=2K | test MSE: 1.2897 -> 1.1429 | Improved |
| use the proposed method instead of target-only training | Doubly debiased robust subsampling · DGP1 covariate shift, Delta=0.0, nS=50K nT=2K | test MSE: 1.0026 -> 0.9918 | Improved |
| use the proposed method instead of IWERM | Doubly debiased robust subsampling · DGP2 label shift, severe, nS=50K nT=2K | misclassification error (%): 24.9731 -> 22.8763 | Improved |
| use leverage-score subsampling instead of uniform | Doubly debiased robust subsampling · DGP4 heavy tails, nu=2, nS=50K nT=2K | test MSE: 2.1047 -> 2.1869 | Hurt |
| increase target sample size nT from 2K to 5K | Doubly debiased robust subsampling · DGP1 covariate shift, Delta=0.4, nS=50K | test MSE (proposed): 1.1429 -> 1.1017 | Improved |
Closest papers by meaning
- Robust Gradient-Based Markov Subsampling (AAAI 2020 · similarity 0.86)
- Robust Sampling for Active Statistical Inference (NeurIPS 2025 · similarity 0.84)
- Relaxing the Accurate Imputation Assumption in Doubly Robust Learning for Debiased Collaborative Filtering (ICML 2024 · similarity 0.83)
- Transfer Learning Meets Functional Linear Regression: No Negative Transfer Under Posterior Drift (AAAI 2025 · similarity 0.83)
- Learning from Noisy Labels via Conditional Distributionally Robust Optimization (NeurIPS 2024 · similarity 0.83)
- Mixed-Sample SGD: an End-to-end Analysis of Supervised Transfer Learning (NeurIPS 2025 · 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.
