Research map / Image generation / GANs
Data-Efficient Instance Generation from Instance Discrimination
Read the paperCheck 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 |
|---|---|---|---|
| Add instance discrimination for real images (ablation) | StyleGAN2-ADA · FFHQ 2K | FID: 15.60 -> 14.15 | Helped |
| Add instance discrimination for real images (ablation) | StyleGAN2-ADA · FFHQ 10K | FID: 7.29 -> 5.98 | Helped |
| Add instance discrimination for real images (ablation) | StyleGAN2-ADA · FFHQ 70K | FID: 3.76 -> 3.56 | Helped |
| Add fake-image instance discrimination without noise perturbation (ablation) | InsGen with real-image instance discrimination · FFHQ 2K and 10K; 70K | FID: 14.15 -> 13.46; 5.98 -> 5.68; 3.56 -> 3.67 | Mixed |
| Add noise perturbation to fake-image instance discrimination (ablation) | InsGen with vanilla fake instance discrimination · FFHQ 2K | FID: 13.46 -> 12.19 | Helped |
| Add noise perturbation to fake-image instance discrimination (ablation) | InsGen with vanilla fake instance discrimination · FFHQ 10K | FID: 5.68 -> 5.30 | Helped |
| Add noise perturbation to fake-image instance discrimination (ablation) | InsGen with vanilla fake instance discrimination · FFHQ 70K | FID: 3.67 -> 3.49 | Helped |
| Add instance-discrimination supervision to the generator (ablation) | InsGen with noise-perturbed fake instance discrimination · FFHQ 2K | FID: 12.19 -> 11.92 | Helped |
| Add instance-discrimination supervision to the generator (ablation) | InsGen with noise-perturbed fake instance discrimination · FFHQ 10K | FID: 5.30 -> 4.90 | Helped |
| Add instance-discrimination supervision to the generator (ablation) | InsGen with noise-perturbed fake instance discrimination · FFHQ 70K | FID: 3.49 -> 3.31 | Helped |
| Use InsGen instead of ADA | StyleGAN2-ADA · FFHQ 2K | FID: 15.60 -> 11.92 | Helped |
| Use InsGen instead of ADA | StyleGAN2-ADA · FFHQ 10K | FID: 7.29 -> 4.90 | Helped |
Closest papers by meaning
- Instance Selection for GANs (NeurIPS 2020 · similarity 0.90)
- Instance-Conditioned GAN (NeurIPS 2021 · similarity 0.90)
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · similarity 0.90)
- Mixed batches and symmetric discriminators for GAN training (ICML 2018 · similarity 0.89)
- Augmentation-Aware Self-Supervision for Data-Efficient GAN Training (NeurIPS 2023 · similarity 0.89)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.88)
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.