AI Paper Atlas

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

Causal Effect Regularization: Automated Detection and Removal of Spurious Correlations

Abhinav Kumar, Amit Deshpande, Amit Sharma

NeurIPS 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Group-robust classification · 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
use Riesz estimator instead of Direct for causal effect (ablation)AutoACER effect estimation · Syn-Text-Obs-Conf, kappa 0.99estimated effect (true 0): 0.22 -> 0.15Improved
use Riesz instead of Direct estimator (ablation)AutoACER effect estimation · MNIST34, kappa 0.8estimated effect (true 0): 0.06 -> 0.04Improved
use Riesz instead of Direct estimator when effect not identified (ablation)AutoACER effect estimation · Syn-Text-Unobs-Conf, kappa 0.99estimated effect (true 0): 0.67 -> 0.70Hurt
use Riesz instead of Direct estimator (ablation)AutoACER effect estimation · MNIST34, kappa 0.99estimated effect (true 0): 0.27 -> 0.31Hurt
raise spurious predictive correlation kappa from 0.5 to 0.99causal effect estimators · Syn-Text-Unobs-Conf, Rieszestimated effect (true 0): 0.06 -> 0.70Hurt
soft causal-effect regularization (AutoACER) instead of hard Mouli+CAD detection and augmentationAutoACER · Syn-Text-Unobs-Confaverage group accuracy and DeltaProbImproved
AutoACER vs baselinesAutoACER · MNIST34DeltaProbImproved
AutoACER vs ERM on accuracyAutoACER · Twitter-AAEaverage group accuracyNo clear effect
Mouli+CAD invariance via counterfactual augmentationMouli+CAD · Twitter-AAEDeltaProbHurt
Mouli+CAD at high correlation (kappa >= 0.8)Mouli+CAD · Syn-Text-Unobs-Confspurious attribute detectionHurt
replace CAD with causal effect regularization in Mouli pipeline (Mouli+AutoACER) (ablation)Mouli · Syn-Text, MNIST34, Twitter-AAEinvariance to spurious attributeImproved
give all methods knowledge of the spurious attribute (ablation)AutoACER and baselines · all datasets (Fig. 7)average group accuracy and DeltaProbNo clear effect

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