AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Counterfactual Residual Data Augmentation for Regression

Hossein Mohebbi, Oliver Schulte, Ke Li, Pascal Poupart

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Data augmentation and generalization bias · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
apply CRDA residual-preserving augmentationMLP regressor · 9 regression datasets (avg)test MSE: average MSE reduction of 22.9%Improved
apply CRDA residual-preserving augmentationXGBoost regressor · 9 regression datasets (avg)test MSE: average MSE reduction of 6.4%Improved
use CRDA instead of CTGAN augmentationMLP · Parkinson's MonitoringMSE change (%): 1454.2 -> -48.2Improved
use CRDA instead of TabDDPM augmentationXGBoost · Parkinson's MonitoringMSE change (%): 286.7 -> -0.3Improved
use CRDA instead of ADA augmentationXGBoost · Synthetic RegressionMSE change (%): 18.3 -> 2.3Improved
use CRDA instead of C-Mixup augmentationXGBoost · Synthetic RegressionMSE change (%): 141.5 -> 2.3Improved
use C-Mixup instead of CRDAXGBoost · Concrete StrengthMSE change (%): -1.7 -> -2.8Hurt
use ADA instead of CRDAMLP · House PriceMSE change (%): -42.3 -> -52.6Hurt
use CRDA on Wine QualityMLP · Wine QualityMSE change (%): -2.9No clear effect
use CRDA with the Wilcoxon-gate fallback on small nXGBoost · Parkinson's Monitoring (n=5875)MSE change (%): 1.65Hurt
use synthetic data with sample size between 2.5k and 20k (ablation)CRDA · Synthetic DGPMSE reduction: between 2.5k and 20k samplesImproved
use very low sample size under 2.5k (ablation)CRDA · Synthetic DGPMSE reduction: <2.5kNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy