Research map / Label-efficient and robust learning / Domain adaptation and generalization
Domain Intersection and Domain Difference
Sagie Benaim, Michael Khaitov, Tomer Galanti, Lior Wolf
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 zero loss (ablation) | Domain intersection/difference model · CelebA Smile to Glasses (loss ablation) | domain-membership classifier accuracy: 85.4% -> 91.8% | Improved |
| add zero loss (ablation) | Domain intersection/difference model · CelebA Glasses to Smile (loss ablation) | domain-membership classifier accuracy: 97.8% -> 99.3% | Improved |
| add adversarial loss (ablation) | Domain intersection/difference model · CelebA Smile to Glasses (loss ablation) | domain-membership classifier accuracy: 64,5% -> 91.8% | Improved |
| add adversarial loss (ablation) | Domain intersection/difference model · CelebA Glasses to Smile (loss ablation) | domain-membership classifier accuracy: 79.3% -> 99.3% | Improved |
| add reconstruction loss (ablation) | Domain intersection/difference model · CelebA Smile to Glasses (loss ablation) | domain-membership classifier accuracy: 50.0% -> 91.8% | Improved |
| add reconstruction loss (ablation) | Domain intersection/difference model · CelebA Glasses to Smile (loss ablation) | domain-membership classifier accuracy: 50.0% -> 99.3% | Improved |
| use symmetric disentangled encoders instead of Guided content transfer | Smile to Glasses translation · CelebA guided translation: Smile to Glasses | domain-membership classifier accuracy: 45.8% -> 91.8% | Improved |
| use ours instead of Fader networks | Facial Hair to Smile translation · CelebA guided translation: Facial Hair to Smile | domain-membership classifier accuracy: 95.4% -> 93.7% | Hurt |
| use ours instead of Vae-NAM | MNIST to SVHN adaptation · MNIST to SVHN | accuracy of mapped label: 51.7% -> 61.0% | Improved |
| use ours instead of Vae-NAM | SVHN to MNIST adaptation · SVHN to MNIST | accuracy of mapped label: 37.4% -> 41.0% | Improved |
Closest papers by meaning
- Domain Separation Networks (NeurIPS 2016 · similarity 0.85)
- Unsupervised Domain Generalization by Learning a Bridge Across Domains (CVPR 2022 · similarity 0.84)
- Bidirectional One-Shot Unsupervised Domain Mapping (ICCV 2019 · similarity 0.84)
- Joint Geometrical and Statistical Alignment for Visual Domain Adaptation (CVPR 2017 · similarity 0.84)
- DDJND: Dual Domain Just Noticeable Difference in Multi-Source Content Images with Structural Discrepancy (AAAI 2025 · similarity 0.82)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.82)
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.
