Research map / Adversarial robustness and security / Attacks on text and graphs
Rob-GAN: Generator, Discriminator, and Adversarial Attacker
Check 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 |
|---|---|---|---|
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · CIFAR10, δmax=0 | accuracy: 81.45% -> 81.1% | Hurt |
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · CIFAR10, δmax=0.02 | accuracy: 69.15% -> 70.41% | Helped |
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · CIFAR10, δmax=0.04 | accuracy: 53.74% -> 57.43% | Helped |
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · CIFAR10, δmax=0.08 | accuracy: 23.58% -> 30.25% | Helped |
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · ImageNet-143 (64px), δmax=0 | accuracy: 20.05% -> 32.4% | Helped |
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · ImageNet-143 (64px), δmax=0.01 | accuracy: 18.3% -> 25.2% | Helped |
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · ImageNet-143 (64px), δmax=0.02 | accuracy: 12.52% -> 19.1% | Helped |
| Add GAN-generated fake-data augmentation to adversarial training | Adversarially trained classifier · ImageNet-143 (64px), δmax=0.03 | accuracy: 8.32% -> 13.7% | Helped |
| Fine-tune the discriminator using the robustness classification loss (ablation) | Rob-GAN discriminator · CIFAR10 and ImageNet-143 (64px) | accuracy | Helped |
| Use the split classification loss instead of the original AC-GAN loss (ablation) | Rob-GAN generator · ImageNet-143 (64px and 128px) | inception score | Helped |
| Use the split classification loss instead of the original AC-GAN loss (ablation) | Rob-GAN generator · Generated image samples | image quality | Helped |
| Add adversarial training to the GAN discriminator (ablation) | GAN training · ImageNet-143 (64px and 128px) | convergence speed: 3\sim 7x | Helped |
Closest papers by meaning
- GLeaD: Improving GANs With a Generator-Leading Task (CVPR 2023 · similarity 0.90)
- A Direct Approach to Robust Deep Learning Using Adversarial Networks (ICLR 2019 · similarity 0.90)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.90)
- MGAN: Training Generative Adversarial Nets with Multiple Generators (ICLR 2018 · similarity 0.90)
- Improving Generalization and Stability of Generative Adversarial Networks (ICLR 2019 · similarity 0.89)
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · 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.