AI Paper Atlas

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

Stable Classification

Dimitris Bertsimas, Jack Dunn, Ivan Paskov

JMLR 2022 · Learning theory and algorithms › Prediction and decision losses · approach: Data attribution and recourse

Read the paper

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 stable cutting-plane training (SCP)SVM · 30 UCI datasetsaccuracy: 0.773 -> 0.795Improved
use stable robust counterpart (SRC)SVM · 30 UCI datasetsaccuracy: 0.773 -> 0.807Improved
use stable robust counterpart (SRC)SVM · 30 UCI datasetsoutput stability: 8.493 × 10−3 -> 5.672 × 10−3Improved
use stable Monte Carlo (SMC)SVM · 30 UCI datasetsoutput stability: 8.493 × 10−3 -> 9.567 × 10−3Hurt
use stable cutting-plane training (SCP)Logistic Regression · 30 UCI datasetsoutput stability: 1.004 × 10−2 -> 8.083 × 10−3Improved
use stable cutting-plane training (SCP)Logistic Regression · 30 UCI datasetsaccuracy: 0.672 -> 0.657Hurt
use stable robust counterpart (SRC)Logistic Regression · 30 UCI datasetsaccuracy: 0.672 -> 0.668No clear effect
use stable Monte Carlo (SMC)Random Forest · 30 UCI datasetsaccuracy: 0.847 -> 0.849No clear effect
use stable cutting-plane training (SCP)Random Forest · 30 UCI datasetsoutput stability: 0.016 -> 0.014Improved
use stable Monte Carlo (SMC)Random Forest · 30 UCI datasetshyperparameter stability: 5.924 -> 5.605Improved
use stable Monte Carlo (SMC)OCT · 30 UCI datasetsoutput stability: 0.029 -> 0.026Improved
use stable training (SMC or SCP)OCT · 30 UCI datasetsaccuracy: 0.828 -> 0.825Hurt

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