Research map / Image generation / GANs
LatentCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directions
Oğuz Kaan Yüksel, Enis Simsar, Ezgi Gülperi Er, Pinar Yanardag
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 |
|---|---|---|---|
| use LatentCLR instead of SeFa for negative Smile direction | StyleGAN2 FFHQ · FFHQ Smile negative | predictor score (lower is stronger removal): 0.50 -> 0.11 | Improved |
| use LatentCLR instead of Ganspace for negative Smile direction | StyleGAN2 FFHQ · FFHQ Smile negative | predictor score: 0.05 -> 0.11 | No clear effect |
| use LatentCLR instead of Ganspace for positive Age direction | StyleGAN2 FFHQ · FFHQ Age positive | predictor score: 0.32 -> 0.42 | Improved |
| use LatentCLR instead of Ganspace for positive Lipstick direction | StyleGAN2 FFHQ · FFHQ Lipstick positive | predictor score: 0.58 -> 0.66 | Improved |
| use LatentCLR instead of Ganspace for positive Smile direction | StyleGAN2 FFHQ · FFHQ Smile positive | predictor score: 0.99 -> 0.99 | No clear effect |
| use LatentCLR instead of Ganspace for negative Age direction | StyleGAN2 FFHQ · FFHQ Age negative | predictor score: 0.23 -> 0.23 | No clear effect |
| use LatentCLR instead of Ganspace for negative Lipstick direction | StyleGAN2 FFHQ · FFHQ Lipstick negative | predictor score: 0.35 -> 0.36 | No clear effect |
| increase the number of directions K beyond that needed (ablation) | LatentCLR · BigGAN/StyleGAN2 | direction diversity | Hurt |
Closest papers by meaning
- Unsupervised Discovery of Interpretable Directions in the GAN Latent Space (ICML 2020 · similarity 0.93)
- Closed-Form Factorization of Latent Semantics in GANs (CVPR 2021 · similarity 0.91)
- NoiseCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directions in Diffusion Models (CVPR 2024 · similarity 0.90)
- Self-Supervised Enhancement of Latent Discovery in GANs (AAAI 2022 · similarity 0.89)
- LARGE: Latent-Based Regression Through GAN Semantics (CVPR 2022 · similarity 0.89)
- Navigating the GAN Parameter Space for Semantic Image Editing (CVPR 2021 · similarity 0.89)
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.
