AI Paper Atlas

Research map / Learning theory and algorithms / Fairness

RISE: Robust Individualized Decision Learning with Sensitive Variables

Xiaoqing Tan, Zhengling Qi, Christopher Warren Seymour, Lu Tang

NeurIPS 2022 · Learning theory and algorithms › Fairness · approach: Fairness auditing and regularization · 3 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 RISE's sensitive-variable-aware infimum objective instead of BaseIndividualized decision rule · Example 1, discrete SObj. (all): 7.03 (0.03) -> 12.0 (0.01)Improved
Use RISE's sensitive-variable-aware quantile objective instead of BaseIndividualized decision rule · Example 1, continuous SObj. (all): 9.12 (0.03) -> 12.2 (0.02)Improved
Use RISE's sensitive-variable-aware infimum objective instead of ExpIndividualized decision rule · Example 2, discrete SObj. (all): 9.12 (0.03) -> 13.5 (0.01)Improved
Use RISE's sensitive-variable-aware quantile objective instead of ExpIndividualized decision rule · Example 2, continuous SObj. (all): 11.1 (0.02) -> 14.1 (0.01)Improved
Use RISE's robust objective instead of the strongest listed baselineNSW job-training decision rule · NSW (log(income+1))Obj. (all): 5.26 (0.04) -> 5.43 (0.04)Improved
Use RISE's robust objective instead of ExpACTG175 treatment decision rule · ACTG175 (CD4 T-cell amount)Obj. (all): 337.5 (1.65) -> 351.5 (1.67)Improved
Use RISE's robust objective instead of BaseSepsis vasopressor decision rule · Sepsis (survival rate)Obj. (all): 0.803 (0.001) -> 0.836 (0.001)Improved
Set the continuous-sensitive-variable quantile level to 0.1 rather than 0.25 (ablation)RISE · Example 2, continuous SObj. (all): 13.8 (0.01) at 0.1; 14.1 (0.01) at 0.25Mixed
Set the continuous-sensitive-variable quantile level to 0.5 rather than 0.25 (ablation)RISE · Example 2, continuous SObj. (all): 17.4 (0.04) at 0.5No clear effect
Make S irrelevant to outcome generation (ablation)RISE versus comparison methods · Example 2, noise-variable simulation, continuous SObj. (all): 27.2 (0.04) for Base, Exp, PT-Base, PT-Exp, and RISENo clear effect

Later papers that compared with this method

3 later accepted papers report RISE as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

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