Research map / Image generation / GANs
Non-Adversarial Image Synthesis With Generative Latent Nearest Neighbors
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 GLANN instead of the strongest listed GAN baseline | GLANN · MNIST | FID: 6.7 ± 0.4 -> 8.6 ± 0.1 | Hurt |
| Use GLANN instead of the strongest listed GAN baseline | GLANN · Fashion | FID: 21.5 ± 1.6 -> 13.0 ± 0.1 | Helped |
| Use GLANN instead of the strongest listed GAN baseline | GLANN · CIFAR10 | FID: 55.2 ± 2.3 -> 46.5 ± 0.2 | Helped |
| Use GLANN instead of the strongest listed GAN baseline | GLANN · CelebA | FID: 38.9 ± 0.9 -> 46.3 ± 0.1 | Hurt |
| Learn a noise-to-latent mapping with IMLE instead of fitting a Gaussian to GLO latent codes (ablation) | GLO · MNIST | F8 / F1/8: (0.845, 0.616) -> (0.971, 0.979) | Helped |
| Learn a noise-to-latent mapping with IMLE instead of fitting a Gaussian to GLO latent codes (ablation) | GLO · Fashion | F8 / F1/8: (0.888, 0.594) -> (0.985, 0.963) | Helped |
| Learn a noise-to-latent mapping with IMLE instead of fitting a Gaussian to GLO latent codes (ablation) | GLO · CIFAR10 | F8 / F1/8: (0.693, 0.680) -> (0.860, 0.825) | Helped |
| Learn a noise-to-latent mapping with IMLE instead of fitting a Gaussian to GLO latent codes (ablation) | GLO · CelebA | F8 / F1/8: (0.509, 0.404) -> (0.574, 0.681) | Helped |
| Use GLANN for fully unsupervised translation rather than GAN-based unconditional generators | GLANN · MNIST → SVHN | translation score: 31.3% | No clear effect |
| Use GLANN for fully unsupervised translation rather than GAN-based unconditional generators | GLANN · SVHN → MNIST | translation score: 25.0% | Hurt |
| Use GLANN for fully unsupervised translation rather than GAN-based unconditional generators | GLANN · Car → Car | translation score: 1.45 | No clear effect |
Closest papers by meaning
- Optimizing the Latent Space of Generative Networks (ICML 2018 · similarity 0.89)
- Large Scale GAN Training for High Fidelity Natural Image Synthesis (ICLR 2019 · similarity 0.89)
- MSG-GAN: Multi-Scale Gradients for Generative Adversarial Networks (CVPR 2020 · similarity 0.89)
- GLeaD: Improving GANs With a Generator-Leading Task (CVPR 2023 · similarity 0.89)
- Few-shot Cross-domain Image Generation via Inference-time Latent-code Learning (ICLR 2023 · similarity 0.89)
- RG-GAN: Dynamic Regenerative Pruning for Data-Efficient Generative Adversarial Networks (AAAI 2024 · similarity 0.87)
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.