Research map / Image generation / GANs
Learning Implicit Generative Models with the Method of Learned Moments
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 |
|---|---|---|---|
| learn moments instead of using random moments (ablation) | MoLM generator · CIFAR-10 | Inception Score: 2.10 -> 6.99 | Helped |
| learn moments instead of using random moments (ablation) | MoLM generator · CIFAR-10 | FID: 160.3 -> 33.8 | Helped |
| learn moments instead of using random moments (ablation) | MoLM generator · CelebA | MS-SSIM: .444 -> .378 | Helped |
| add gradient features to hidden-unit activations (ablation) | MoLM moment features on CIFAR-10 · CIFAR-10 | Inception Score and FID | Helped |
| use activations alone instead of gradient features (ablation) | MoLM moment features on CIFAR-10 · CIFAR-10 | sample quality | Hurt |
| use gradient features alone rather than gradient and hidden-unit features (ablation) | MoLM moment features on CIFAR-10 · CIFAR-10 | sample quality | Mixed |
| use MoLM rather than SN-GAN | Conv. generator · CIFAR-10 | FID-5K/50K: 25.5/- -> 25.0/20.3 | Helped |
| increase MoLM moment network size from 1024 to 1536 (ablation) | MoLM with Conv. generator · CIFAR-10 | Inception Score: 7.55 ± .08 -> 7.90 ± .10 | Helped |
| increase MoLM moment network size from 1024 to 1536 (ablation) | MoLM with Conv. generator · CIFAR-10 | FID-5K/50K: 25.0/20.3 -> 23.3/18.9 | Helped |
Closest papers by meaning
- Improving Generative Moment Matching Networks with Distribution Partition (AAAI 2021 · similarity 0.85)
- Inductive Moment Matching (ICML 2025 · similarity 0.84)
- MCA: Moment Channel Attention Networks (AAAI 2024 · similarity 0.84)
- Exact Gaussian Moment Matching for Residual Networks: a Second-Order Method (NeurIPS 2026 · similarity 0.84)
- Moment- and Power-Spectrum-Based Gaussianity Regularization for Text-to-Image Models (NeurIPS 2025 · similarity 0.83)
- Generative Video Diffusion for Unseen Novel Semantic Video Moment Retrieval (AAAI 2025 · similarity 0.83)
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.