Research map / Image generation / GANs
ClusterGAN: Latent Space Clustering in 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 a discrete-continuous latent prior, inverse network, and clustering-specific cycle loss rather than the strongest listed baseline | ClusterGAN · Fashion-10 | ACC: 0.61 -> 0.63 | Helped |
| Use a discrete-continuous latent prior, inverse network, and clustering-specific cycle loss rather than the strongest listed baseline | ClusterGAN · 10x_73k | ACC: 0.71 -> 0.81 | Helped |
| Use a discrete-continuous latent prior, inverse network, and clustering-specific cycle loss rather than the strongest listed baseline | ClusterGAN · Pendigits | ACC: 0.72 -> 0.77 | Helped |
| Use a discrete-continuous latent prior, inverse network, and clustering-specific cycle loss rather than the strongest listed baseline | ClusterGAN · MNIST | NMI: 0.90 -> 0.89 | Hurt |
| Replace the nonlinear generator and Gaussian latent prior with a linear generator and one-hot encoded normal latent variables (ablation) | GAN latent-space clustering on the 10-Gaussian toy data · Synthetic mixture of 10 Gaussians in R^100 | ACC: 0.73 -> 1.0 | Helped |
| Set MNIST cluster count K to 7 instead of 10 (ablation) | ClusterGAN · MNIST | ACC: 0.95 -> 0.72 | Hurt |
| Set MNIST cluster count K to 9 instead of 10 (ablation) | ClusterGAN · MNIST | ACC: 0.95 -> 0.80 | Hurt |
| Set MNIST cluster count K to 11 instead of 10 (ablation) | ClusterGAN · MNIST | ACC: 0.95 -> 0.90 | Hurt |
| Set MNIST cluster count K to 13 instead of 10 (ablation) | ClusterGAN · MNIST | ACC: 0.95 -> 0.84 | Hurt |
Closest papers by meaning
- Balanced Self-Paced Learning for Generative Adversarial Clustering Network (CVPR 2019 · similarity 0.91)
- Top-Down Deep Clustering with Multi-Generator GANs (AAAI 2022 · similarity 0.91)
- Unveiling the Latent Space Geometry of Push-Forward Generative Models (ICML 2023 · similarity 0.90)
- Complexity Matters: Rethinking the Latent Space for Generative Modeling (NeurIPS 2023 · similarity 0.89)
- Exploring and Exploiting Hubness Priors for High-Quality GAN Latent Sampling (ICML 2022 · similarity 0.89)
- Stein Latent Optimization for Generative Adversarial Networks (ICLR 2022 · similarity 0.89)
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.