AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

SPARC: Scalable Path-Specific Counterfactual Fairness via Causal Conditional Independence

Bowei Tian, Yexiao He, Ziyao Wang, Meng Liu, Yongkai Wu, Ang Li

ECCV 2026 · Probabilistic and scientific modeling › Causal discovery · approach: Causal approaches to fairness · 1 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 SPARC instead of vanilla fine-tuningResNet50 · TCGA-LUADEOd (x1e-3): 7.21 -> 1.46Improved
use SPARC instead of vanilla fine-tuningResNet50 · MIMIC-CXRAccuracy (%): 85.46 -> 84.26Hurt
use SPARC instead of Group-DROResNet50 · CheXpertEOd (x1e-3): 3.07 -> 1.20Improved
use ViT backbone instead of ResNet50 (ablation)SPARC · MIMIC-CXRAccuracy (%): 84.26 -> 86.95Improved
use ViT backbone instead of ResNet50 (ablation)SPARC · MIMIC-CXREOp (x1e-2): 2.02 -> 3.35Hurt
use age (binarized at 60) as sensitive attribute (ablation)SPARC · MIMIC-CXREOd (x1e-3): 9.38 -> 1.45Improved
increase entropy loss weight alpha from 0 to 4 (ablation)SPARC · MIMICEOd (x1e-3): 2.32 -> 1.13Improved
increase entropy loss weight alpha from 0 to 4 (ablation)SPARC · TCGAEOd (x1e-3): 4.79 -> 1.46Improved
increase generator loss weight beta from 0 to 4 (ablation)SPARC · MIMICAccuracy (%): 80.69 -> 87.79Improved
increase generator loss weight beta from 0 to 4 (ablation)SPARC · MIMICEOd (x1e-3): 0.98 -> 5.50Hurt
increase mask noise strength eta (ablation)SPARC · MIMIC-CXRAccuracy (%)Hurt
use SPARC instead of AD on synthetic SCMSPARC · Synthetic SCM|IE(S)| preserved: 0.061 -> 0.171Improved

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