Research map / Label-efficient and robust learning / Recognition and long-tail
Robust Correction of Sampling Bias using Cumulative Distribution Functions
Bijan Mazaheri, Siddharth Jain, Jehoshua Bruck
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 the CDF-based V-matrix instead of importance-weighting baselines | SVM · twonorm, replicating [cite] | normalized error: 0.939(0.236) -> 0.935( 0.118) | Improved |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · ringnorm, replicating [cite] | normalized error: 0.873 (.244) -> 0.967( 0.084 ) | Hurt |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · cancer, single feature bias | normalized error: 1.019(0.107) -> 1.072(0.122) | Hurt |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · diabetes, single feature bias | normalized error: 1.002(0.017) -> 0.994(0.054) | Improved |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · banknote, single feature bias | normalized error: 0.995(0.021) -> 0.951(0.242) | Improved |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · ringnorm, single feature bias | normalized error: 1.000(0.004) -> 0.905(0.086) | Improved |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · twonorm, single feature bias | normalized error: 0.952(0.186) -> 1.187(0.200) | Hurt |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · cancer, multiple feature bias | normalized error: 1.015(0.057) -> 1.075(0.115) | Hurt |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · diabetes, multiple feature bias | normalized error: 0.997(0.014) -> 1.006(0.039) | Hurt |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · banknote, multiple feature bias | normalized error: 1.008(0.110) -> 0.919(0.216) | Improved |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · ringnorm, multiple feature bias | normalized error: 0.978(0.032) -> 0.923(0.081) | Improved |
| Use the CDF-based V-matrix instead of the strongest baseline | SVM · twonorm, multiple feature bias | normalized error: 0.959(0.203) -> 1.247(0.277) | Hurt |
Closest papers by meaning
- Not all distributional shifts are equal: Fine-grained robust conformal inference (ICML 2024 · similarity 0.86)
- Quantifying Uncertainty in the Presence of Distribution Shifts (NeurIPS 2025 · similarity 0.86)
- Minimax Optimal Two-Stage Algorithm For Moment Estimation Under Covariate Shift (ICLR 2025 · similarity 0.85)
- Computational Efficiency under Covariate Shift in Kernel Ridge Regression (NeurIPS 2025 · similarity 0.85)
- Beyond the Training Distribution: Evaluating Predictions Under Distribution Shift and Selection Bias (NeurIPS 2026 · similarity 0.85)
- Statistical Inference Under Constrained Selection Bias (ICML 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.
