Research map / Image generation / GANs
MisGAN: Learning from Incomplete Data with Generative Adversarial Networks
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 |
|---|---|---|---|
| Use convolutional networks instead of fully connected networks | MisGAN · MNIST, non-overlapping square quadrants | FID: 20.6 -> 10.8 | Helped |
| Add an explicit mask discriminator and train the mask distribution (ablation) | MisGAN · MNIST missing-data modeling | mask and data distribution quality | Helped |
| Set the mask-generator data-loss coefficient to a small positive value, such as 0.2, instead of zero (ablation) | MisGAN · MNIST missing-data modeling | performance | Helped |
| Use fully connected rather than convolutional networks (ablation) | MisGAN · MNIST, independent dropout at high missing rates | FID | Hurt |
| Remove the mask generator and discriminator while jointly training the imputer and data generator (ablation) | MisGAN imputer · MNIST imputation | imputation quality | Hurt |
| Use independent dropout rather than square observations at the same missing rate | MisGAN · MNIST data generation | FID | Hurt |
| Use MisGAN instead of the tested baseline methods for imputation | MisGAN imputer · MNIST, CIFAR-10, and CelebA; square observation and independent dropout | FID | Helped |
| Train GAIN longer beyond its best-performing epoch | GAIN · Block-missing imputation | imputation quality | Hurt |
Closest papers by meaning
- GAIN: Missing Data Imputation using Generative Adversarial Nets (ICML 2018 · similarity 0.92)
- GAMIN: Generative Adversarial Multiple Imputation Network for Highly Missing Data (CVPR 2020 · similarity 0.89)
- CollaGAN: Collaborative GAN for Missing Image Data Imputation (CVPR 2019 · similarity 0.89)
- No Modes Left Behind: Capturing the Data Distribution Effectively Using GANs (AAAI 2018 · similarity 0.89)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.87)
- Missing Data Imputation by Reducing Mutual Information with Rectified Flows (NeurIPS 2025 · similarity 0.86)
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.