AI Paper Atlas

Research map / Learning theory and algorithms / Fairness

Fairness without Demographics through Adversarially Reweighted Learning

Preethi Lahoti, Alex Beutel, Jilin Chen, Kang Wook Lee, Flavien Prost, Nithum Thain, Xuezhi Wang, Ed Chi

NeurIPS 2020 · Learning theory and algorithms › Fairness · approach: Post-processing and transport for fairness

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 adversarial reweighting based on computationally identifiable errorsBinary classification without observed demographics · AdultAUC (min): 0.867 -> 0.881Improved
Use adversarial reweighting based on computationally identifiable errorsBinary classification without observed demographics · LSACAUC (min): 0.790 -> 0.798Improved
Use adversarial reweighting based on computationally identifiable errorsBinary classification without observed demographics · COMPASAUC (min): 0.674 -> 0.658No clear effect
Use adversarial reweighting based on computationally identifiable errorsBinary classification without observed demographics · AdultAUC (minority): 0.875 -> 0.942Improved
Use adversarial reweighting based on computationally identifiable errorsBinary classification without observed demographics · LSACAUC (minority): 0.824 -> 0.832Improved
Use adversarial reweighting based on computationally identifiable errorsBinary classification without observed demographics · COMPASAUC (minority): 0.774 -> 0.785Improved
Use adversarially learned weights rather than group-based inverse probability weightsFair classification · LSACAUC (min): IPW(S) 0.772; ARL 0.798Improved
Include class label Y as adversary input (ablation)ARL with protected features S available · AdultAUC (min): 0.875 -> 0.882Improved
Include class label Y as adversary input (ablation)ARL with protected features S available · LSACAUC (min): 0.799 -> 0.801Improved
Include class label Y as adversary input (ablation)ARL with protected features S available · COMPASAUC (min): 0.675 -> 0.681Improved
Add protected features S to the adversary input X+Y (ablation)ARL · AdultAUC (min): 0.881 -> 0.881No clear effect
Add protected features S to the adversary input X+Y (ablation)ARL · LSACAUC (min): 0.798 -> 0.808Improved

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