AI Paper Atlas

Research map / Learning theory and algorithms / Fairness

FairSplit: Decomposing the Embedding Space for Fair Classification

Andrii Shkabrii, Anna Beer, Ylli Sadikaj, Claudia Plant

NeurIPS 2026 · Learning theory and algorithms › Fairness · approach: Fairness-constrained classification · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes FairSplit, which decomposes the embedding space into a Fair Space used for prediction and a Sensitive Space that isolates sensitive attributes for fair classification.

Problem

Fair classification must keep high accuracy while limiting reliance on sensitive information in learned representations.

Approach

Learns a decomposition of the embedding space into a Fair Space used for prediction and a Sensitive Space that isolates sensitive information, trained jointly with the classification task.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
learn a mask-based Fair Space / Sensitive Space decomposition instead of applying HSIC to the full embedding (ablation)FairSplit · CelebA-AΔDP: 26.6 -> 18.3Improved
add HSIC regularization on the embedding without decomposition (ablation)ERM · CelebA-AΔDP: 52.8 -> 26.6Improved
add the learned mask before any loss acts on the Sensitive Space (ablation)FairSplit · CelebA-AΔDPImproved
regularize the Sensitive Space (lambda_s, rho_s > 0) so sensitive information moves into it (ablation)FairSplit · CelebA-AΔDPImproved
increase rho_s to shrink Fair Space dimensionality (ablation)FairSplit · CelebA-AΔDPImproved
increase the independence weight rho_f (ablation)FairSplit · CelebA-AΔDPImproved
vary the Fair Space compression weight lambda_f (ablation)FairSplit · CelebA-AΔDPNo clear effect
change HSIC kernel type or bandwidth (ablation)FairSplit · CelebA-AΔDPNo clear effect
use FairSplit instead of ERM when fine-tuning ViT-B/16ViT-B/16 · FairFace age>=30ΔDP: 23.0 -> 9.5Improved
use FairSplit instead of ERM when fine-tuning ViT-B/16ViT-B/16 · FairFace age>=30AP: 88.9 -> 86.4Hurt
transfer a FairSplit Fair Space to a new linear head on frozen CLIP ViT-L/14CLIP ViT-L/14 (frozen) · CelebA ageΔDP: 21.4 -> 9.1Improved
30% of sensitive labels missing during training (ablation)FairSplit · CelebA-AΔDP: 52.8 (ERM) vs 28.0Mixed

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