Research map / Adversarial robustness and security / Certified and empirical robustness
GAMIN: Generative Adversarial Multiple Imputation Network for Highly Missing Data
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 GAMIN instead of misGAN | MNIST imputation · MNIST dropout missing 80%–95% | RMSE: 0.255 -> 0.224; 0.282 -> 0.251; 0.324 -> 0.272; 0.368 -> 0.306 | Helped |
| use GAMIN instead of misGAN | MNIST imputation · MNIST dropout missing 90% | top-3 classification accuracy: 56.9 -> 66.9 | Helped |
| use GAMIN instead of misGAN | CelebA imputation · CelebA dropout missing 90% | RMSE: 0.129 -> 0.115 | Helped |
| add candidate-imputation substitution instead of uniform substitution (ablation) | GAMIN · MNIST dropout missing 90% | RMSE: 0.280 -> 0.272 | Helped |
| add confidence-related training losses (ablation) | GAMIN · MNIST dropout missing 90% | RMSE: 0.275 -> 0.272 | Helped |
| add test-time confidence prediction and ranking (ablation) | GAMIN · MNIST dropout missing 90% | RMSE: 0.304 -> 0.272 | Helped |
| use conditional substitution (ablation) | GAMIN · MNIST dropout missing 90% | RMSE: 0.272 -> 0.270 | Helped |
| use conditional substitution (ablation) | GAMIN · MNIST dropout missing 90% | top-3 classification accuracy: 66.9 -> 71.6 | Helped |
| use GAMIN instead of GAIN | MNIST imputation · MNIST dropout missing 50% | RMSE: 0.130 -> 0.140 | Hurt |
| use GAMIN instead of GAIN | MNIST imputation · MNIST dropout missing 50% | top-3 classification accuracy: 98.0 -> 98.2 | Helped |
Closest papers by meaning
- Meta-GAIN for Missing Data Imputation (AAAI 2026 · similarity 0.93)
- GAIN: Missing Data Imputation using Generative Adversarial Nets (ICML 2018 · similarity 0.92)
- CollaGAN: Collaborative GAN for Missing Image Data Imputation (CVPR 2019 · similarity 0.91)
- MisGAN: Learning from Incomplete Data with Generative Adversarial Networks (ICLR 2019 · similarity 0.89)
- Missing Data Imputation by Reducing Mutual Information with Rectified Flows (NeurIPS 2025 · similarity 0.89)
- DiffPuter: Empowering Diffusion Models for Missing Data Imputation (ICLR 2025 · similarity 0.85)
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.