AI Paper Atlas

Research map / Learning theory and algorithms / Fairness

Subgroup Robustness Grows On Trees: An Empirical Baseline Investigation

Joshua P Gardner, Zoran Popovi, Ludwig Schmidt

NeurIPS 2022 · Learning theory and algorithms › Fairness · approach: Fairness-constrained classification · 5 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 tree-based models (GBM, LightGBM, XGBoost, Random Forest) instead of robust or fairness-enhancing modelstabular subgroup robustness · eight tabular datasetsoverall, worst-group accuracy and accuracy disparityImproved
use XGBoost instead of Group DRO and chi^2 DROXGBoost · ACS Income, ACS Public Coverage, LARCaccuracy and worst-group accuracyImproved
use DORO CVaR instead of tree modelsrobust models · Adultmaximum worst-group accuracyNo clear effect
use Inprocessing instead of tree modelsfairness-enhancing models · Communities and Crimemaximum worst-group accuracyNo clear effect
select model by best CVaR instead of best accuracy (ablation)DRO chi^2 and Group DRO · ACS Income and other datasetsworst-group accuracyHurt
select model by best CVaR instead of best accuracy (ablation)tree-based models · ACS Income and other datasetsworst-group accuracyNo clear effect
use XGBoost instead of DRO chi^2 for hyperparameter sweep costtree models · ACS Incomesweep compute cost: 58 -> 1Improved
use tree-based models instead of others for hyperparameter sensitivitytree models · all datasetsworst-group accuracy decline over ranked configurationsImproved
pair complementary metrics for model selectionall models · all datasetsmetric correlation: 0.87, 0.99, 0.79Improved
select by robust risk (CVaR DORO), evaluate by accuracyrobust models · all datasetsaccuracyHurt

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