Research map / Label-efficient and robust learning / Recognition and long-tail
Activation Maximization Generative Adversarial Nets
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 |
|---|---|---|---|
| Replace the overall real-class target with an explicit target class for each generated sample | LabelGAN, yielding AM-GAN · CIFAR-10 | Inception Score: 8.63 ± 0.08 -> 8.83 ± 0.09 | Helped |
| Replace the overall real-class target with an explicit target class for each generated sample | LabelGAN, yielding AM-GAN · CIFAR-10 | AM Score: 0.13 ± 0.00 -> 0.08 ± 0.00 | Helped |
| Add adversarial training to the auxiliary classifier (ablation) | AC-GAN* · CIFAR-10 | Inception Score: 7.41 ± 0.09 -> 8.56 ± 0.11 | Helped |
| Add adversarial training to the auxiliary classifier (ablation) | AC-GAN* · CIFAR-10 | AM Score: 0.17 ± 0.00 -> 0.10 ± 0.00 | Helped |
| Use dynamic labeling instead of predefined labels (ablation) | AM-GAN · CIFAR-10 | Inception Score: 8.35 ± 0.12 -> 8.83 ± 0.09 | Helped |
| Use dynamic labeling instead of predefined labels (ablation) | AM-GAN · CIFAR-10 | AM Score: 0.05 ± 0.00 -> 0.08 ± 0.00 | Hurt |
| Use dynamic labeling instead of predefined labels (ablation) | AM-GAN · Tiny ImageNet | Inception Score: 9.55 ± 0.11 -> 11.45 ± 0.15 | Helped |
| Use dynamic labeling instead of predefined labels (ablation) | AM-GAN · Tiny ImageNet | AM Score: 0.61 ± 0.01 -> 0.88 ± 0.01 | Hurt |
| Restore the omitted fake-sample auxiliary-classifier loss (ablation) | AC-GAN* · CIFAR-10, dynamic labels | Inception Score: 7.41 to 6.48 | Hurt |
| Restore the omitted fake-sample auxiliary-classifier loss (ablation) | AC-GAN* · CIFAR-10, dynamic labels | AM Score: 0.17 to 0.43 | Hurt |
| Reduce the generator's auxiliary-classifier loss weight to one tenth of the GAN loss (ablation) | AC-GAN* · CIFAR-10 | Inception Score: 7.19 | Hurt |
Closest papers by meaning
- An Adversarial Framework for Generating Unseen Images by Activation Maximization (AAAI 2022 · similarity 0.90)
- Improved Techniques for Training GANs (NeurIPS 2016 · similarity 0.89)
- MGAN: Training Generative Adversarial Nets with Multiple Generators (ICLR 2018 · similarity 0.89)
- On-Line Adaptative Curriculum Learning for GANs (AAAI 2019 · similarity 0.89)
- Active Generative Adversarial Network for Image Classification (AAAI 2019 · similarity 0.89)
- Improving the Training of Data-Efficient GANs via Quality Aware Dynamic Discriminator Rejection Sampling (CVPR 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.