Research map / Label-efficient and robust learning / Recognition and long-tail
Generative Imputation and Stochastic Prediction
Mohammad Kachuee, Kimmo Kärkkäinen, Orpaz Goldstein, Sajad Darabi, Majid Sarrafzadeh
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 stochastic predictor with N=128 imputations instead of N=1 (ablation) | GI · CIFAR-10 rectangular 90% missing | top-1 accuracy: 29.4 -> 32.1 | Improved |
| use stochastic predictor with N=128 instead of N=1 (ablation) | GI · CIFAR-10 rectangular 20% missing | top-1 accuracy: 83.6 -> 84.0 | Improved |
| use stochastic predictor with N=128 instead of N=1 (ablation) | GI · CIFAR-10 rectangular 70% missing | top-1 accuracy: 56.7 -> 59.1 | Improved |
| increase ensemble size N (ablation) | GI · Landsat | top-1 accuracy | Improved |
| add self-attention layers before each residual block (ablation) | GI · CIFAR-10 rectangular missingness | FID | Improved |
| add self-attention layers before each residual block (ablation) | GI · CIFAR-10 uniform missingness | FID | No clear effect |
| add Gaussian training noise std 0.0125 (ablation) | GI · CIFAR-10 uniform 40% missing | top-1 accuracy: 87.1 -> 87.3 | Improved |
| add Gaussian training noise std 0.0125 (ablation) | GI · CIFAR-10 rectangular 40% missing | top-1 accuracy: 76.9 -> 76.8 | Hurt |
| add Gaussian training noise std 0.1 (ablation) | GI · CIFAR-10 uniform 40% missing | top-1 accuracy: 87.1 -> 82.0 | Hurt |
| add discriminator hint vector (ablation) | GI · CIFAR-10 uniform missingness | imputation quality | No clear effect |
| add discriminator hint vector (ablation) | GI · CIFAR-10 rectangular missingness | training stability | Hurt |
| increase weight lambda of MSE loss term in imputer (ablation) | GAIN · synthetic Gaussian | sample variance | Hurt |
Closest papers by meaning
- From Predictive Methods to Missing Data Imputation: An Optimization Approach (JMLR 2018 · similarity 0.88)
- Improving the imputation of missing data with Markov Blanket discovery (ICLR 2023 · similarity 0.88)
- Distribution Shift in Missing Data Imputation: A Risk-Based Perspective and Importance-Weighted Correction under MAR (NeurIPS 2026 · similarity 0.87)
- Random features models: a way to study the success of naive imputation (ICML 2024 · similarity 0.87)
- Learning What Not to Impute: An Uncertainty-Aware Diffusion Framework for Meaningful Missingness (NeurIPS 2026 · similarity 0.86)
- Prediction models that learn to avoid missing values (ICML 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.
