AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

On Human-Aligned Risk Minimization

Liu Leqi, Adarsh Prasad, Pradeep K Ravikumar

NeurIPS 2019 · Learning theory and algorithms › Prediction and decision losses · approach: Losses for decision-making · 13 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
minimize human risk (CPT-weighted, b=.3) instead of expected lossERM · synthetic regression, minority groupminority riskImproved
use EHRM instead of ERMERM · synthetic regression, population/majoritymajority riskHurt
increase b toward 1 (weighting becomes identity) (ablation)EHRM · synthetic regressionminority riskHurt
use EHRM instead of ERMERM logistic regression · COMPAS recidivismmisclassification rate / worst-subgroup gapMixed
use EHRM instead of ERMERM · UTKFace gender classificationaccuracy: .8767 ±.0060 -> .8751 ±.0052Hurt
use EHRM instead of ERMERM · UTKFace gender classificationstatistical parity difference: -.0881 ±.0208 -> -.0825 ±.0220Improved
use EHRM instead of ERMERM · UTKFace gender classificationTheil index: .0855 ±.0071 -> .0859 ±.0058No clear effect
use small CVaR alpha instead of EHRMEHRM · synthetic regressionminority vs majority riskMixed

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