Research map / Image generation / GANs
Energy-based 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 |
|---|---|---|---|
| Place a bottom-layer-cost Ladder Network in the EBGAN framework | Ladder Network · PI-MNIST, 100 labels | error rate (%): 1.36$\pm$0.21 -> 1.04$\pm$0.12 | Helped |
| Place a bottom-layer-cost Ladder Network in the EBGAN framework | Ladder Network · PI-MNIST, 200 labels | error rate (%): 1.24$\pm$0.09 -> 0.99$\pm$0.12 | Helped |
| Place a bottom-layer-cost Ladder Network in the EBGAN framework | Ladder Network · PI-MNIST, 1000 labels | error rate (%): 1.04$\pm$0.06 -> 0.89$\pm$0.04 | Helped |
| Use the EBGAN auto-encoder model instead of a probabilistic GAN discriminator | GAN training on MNIST · MNIST | modified inception score I' | Mixed |
| Add the Pulling-away Term (PT) to the generator loss (ablation) | EBGAN auto-encoder · MNIST | generation quality | Mixed |
| Decay the margin from 16 to 0 during early training | EBGAN-LN · PI-MNIST | error rate (%): 1.36$\pm$0.21 -> 1.04$\pm$0.12; 1.24$\pm$0.09 -> 0.99$\pm$0.12; 1.04$\pm$0.06 -> 0.89$\pm$0.04 | Helped |
| Use a non-decaying margin rather than decaying it toward 0 | EBGAN-LN · PI-MNIST | semi-supervised learning performance | Hurt |
| Set the energy margin too large | EBGAN · MNIST generation | training stability | Hurt |
| Set the energy margin too small | EBGAN · MNIST generation | mode coverage | Hurt |
Closest papers by meaning
- Calibrating Energy-based Generative Adversarial Networks (ICLR 2017 · similarity 0.92)
- Your GAN is Secretly an Energy-based Model and You Should Use Discriminator Driven Latent Sampling (NeurIPS 2020 · similarity 0.88)
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · similarity 0.88)
- Scalable Energy-Based Models via Adversarial Training: Unifying Discrimination and Generation (ICLR 2026 · similarity 0.88)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.87)
- Improving Adversarial Energy-Based Model via Diffusion Process (ICML 2024 · 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.