Research map / Label-efficient and robust learning / Recognition and long-tail
Fairness by Learning Orthogonal Disentangled Representations
Mhd Hasan Sarhan, Nassir Navab, Abouzar Eslami, Shadi Albarqouni
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add orthogonal-disentangled representation learning vs plain classifier | ResNet-18 classifier · CIFAR-10 | sensitive accuracy: 0.2344 -> 0.1907 | Improved |
| add orthogonal-disentangled representation learning vs plain classifier | ResNet-18 classifier · CIFAR-100 | sensitive accuracy: 0.3069 -> 0.1447 | Improved |
| add orthogonal-disentangled representation learning vs plain classifier | ResNet-18 classifier · CIFAR-10 | target accuracy: 0.9775 -> 0.9725 | Hurt |
| add orthogonal-disentangled representation learning vs plain classifier | ResNet-18 classifier · CIFAR-100 | target accuracy: 0.7199 -> 0.7074 | Hurt |
| add orthogonal-disentangled learning vs Roy et al. at comparable target accuracy | Roy et al. (trade-off #2) · CIFAR-100 | sensitive accuracy: 0.1643 -> 0.1447 | Improved |
| add entropy loss and KL with orthogonal means (full model) vs plain classifier (ablation) | baseline classifier · five datasets | sensitive accuracy | Improved |
| add KL orthogonal loss without entropy loss (ablation) | KL Orth. w/o Entropy · CIFAR-10 | sensitive accuracy: 0.26% drop in target performance | Improved |
| add entropy loss alone or KL loss alone vs multi-task learning (ablation) | multi-task learning · ablation datasets | sensitive accuracy | No clear effect |
| add auxiliary sensitive classifier (multi-task) to the target classifier (ablation) | baseline classifier · three of five datasets | sensitive accuracy | Improved |
| add orthogonality to disentangled KL loss (orthogonal means instead of similar means) (ablation) | Entropy + KL · ablation datasets | sensitive accuracy | Improved |
Closest papers by meaning
- On the Fairness of Disentangled Representations (NeurIPS 2019 · similarity 0.92)
- Flexibly Fair Representation Learning by Disentanglement (ICML 2019 · similarity 0.91)
- FADES: Fair Disentanglement with Sensitive Relevance (CVPR 2024 · similarity 0.87)
- Disentangled Representation Learning with the Gromov-Monge Gap (ICLR 2025 · similarity 0.87)
- Disentangled Representation Learning via Modular Compositional Bias (NeurIPS 2025 · similarity 0.86)
- $\alpha$TC-VAE: On the relationship between Disentanglement and Diversity (ICLR 2024 · similarity 0.86)
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.
