AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Fairness without Demographics through Knowledge Distillation

Junyi Chai, Taeuk Jang, Xiaoqian Wang

NeurIPS 2022 · Label-efficient and robust learning › Knowledge distillation · approach: Self and data-free distillation · 4 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
replace hard labels with softmax soft labels from teacherResNet-18 student · COMPAS raceequalized odds: 38.34±3.37% -> 21.32±1.97%Improved
replace hard labels with linear soft labels from teacherResNet-18 student · CelebA gender classification (age)equalized odds: 17.32±1.78% -> 9.64±2.14%Improved
replace hard labels with softmax soft labelsResNet-18 student · New Adult raceequalized odds: 14.43±1.24% -> 10.34±1.22%Improved
replace hard labels with softmax soft labelsResNet-18 student · CelebA attractiveness (gender)accuracy: 81.67±0.63% -> 80.87±0.14%Hurt
replace hard labels with softmax soft labelsResNet-18 student · COMPAS raceaccuracy: 64.13±0.32% -> 63.47±0.44%Hurt
use soft labels (binary to linear) in preliminary label experiment (ablation)student on new Adult · New Adult raceequalized odds: 13.22±1.70% -> 10.13±1.06%Improved
assign random labels instead of teacher soft labels (ablation)student on new Adult · New Adult raceequalized odds: 13.22±1.70% -> 13.15±1.12%No clear effect
add DRO (Max-Min fairness) trainingResNet-18 student · CelebA gender classification (age)accuracy: 90.43±0.23% -> 74.46±0.37%Hurt
add DRO (Max-Min fairness) trainingResNet-18 student · COMPAS raceequalized odds: 38.34±3.37% -> 30.43±3.24%Mixed
add ARL adversarial reweightingResNet-18 student · New Adult genderequalized odds: 18.82±1.82% -> 16.52±1.96%No clear effect
add FairRFResNet-18 student · New Adult genderequalized odds: 18.82±1.82% -> 13.54±1.26%Improved
use large overfitted teacher model directly instead of studentResNet-152 teacher · COMPAS raceequalized odds: 38.34±3.37% -> 39.64%Hurt

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