Research map / Image generation / GANs
Memorization Precedes Generation: Learning Unsupervised GANs with Memory 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 MemoryGAN instead of Fisher GAN | unsupervised GAN image generation · CIFAR10 | inception score: 7.90 ± 0.05 -> 8.04 ± 0.13 | Helped |
| add incremental EM memory updates (ablation) | MemoryGAN · CIFAR10 | inception score: 6.67 ± 0.17 -> 8.04 ± 0.13 | Helped |
| add incremental EM memory updates (ablation) | MemoryGAN · affine-MNIST | inception score: 8.17 ± 0.04 -> 8.60 ± 0.02 | Helped |
| add incremental EM memory updates (ablation) | MemoryGAN · Fashion-MNIST | inception score: 6.05 ± 0.28 -> 6.39 ± 0.29 | Helped |
| add memory-slot sampling to the generator (ablation) | MemoryGAN · CIFAR10 | inception score: 6.39 ± 0.02 -> 8.04 ± 0.13 | Helped |
| add memory-slot sampling to the generator (ablation) | MemoryGAN · affine-MNIST | inception score: 8.11 ± 0.03 -> 8.60 ± 0.02 | Helped |
| add memory-slot sampling to the generator (ablation) | MemoryGAN · Fashion-MNIST | inception score: 6.02 ± 0.27 -> 6.39 ± 0.29 | Helped |
| add the memory network (ablation) | MemoryGAN · CIFAR10 | inception score: 5.35 ± 0.17 -> 8.04 ± 0.13 | Helped |
| add the memory network (ablation) | MemoryGAN · affine-MNIST | inception score: 8.00 ± 0.01 -> 8.60 ± 0.02 | Helped |
| add the memory network (ablation) | MemoryGAN · Fashion-MNIST | inception score: 5.82 ± 0.19 -> 6.39 ± 0.29 | Helped |
Closest papers by meaning
- Memory Replay GANs: Learning to Generate New Categories without Forgetting (NeurIPS 2018 · similarity 0.90)
- Detecting Overfitting of Deep Generative Networks via Latent Recovery (CVPR 2019 · similarity 0.90)
- Re-GAN: Data-Efficient GANs Training via Architectural Reconfiguration (CVPR 2023 · similarity 0.89)
- When, Why, and Which Pretrained GANs Are Useful? (ICLR 2022 · similarity 0.89)
- Sequential Training of GANs Against GAN-Classifiers Reveals Correlated "Knowledge Gaps" Present Among Independently Trained GAN Instances (CVPR 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.