Research map / Label-efficient and robust learning / Semi-supervised learning
Semi-supervised Learning with GANs: Manifold Invariance with Improved Inference
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 |
|---|---|---|---|
| Add GAN-estimated tangent invariance (ablation) | FM-GAN · SVHN (500 labeled examples) | test error (%): 18.44±4.8 -> 5.88±1.5 | Helped |
| Add GAN-estimated tangent invariance (ablation) | FM-GAN · SVHN (1000 labeled examples) | test error (%): 8.11±1.3 -> 5.26±1.1 | Helped |
| Add GAN-estimated tangent invariance (ablation) | FM-GAN · CIFAR-10 (1000 labeled examples) | test error (%): 21.83±2.01 -> 20.23±1.3 | Helped |
| Add GAN-estimated tangent invariance (ablation) | FM-GAN · CIFAR-10 (4000 labeled examples) | test error (%): 18.63 ± 2.32 -> 16.96 ± 1.4 | Helped |
| Add Jacobian regularization (ablation) | FM-GAN · SVHN (500 labeled examples) | test error (%): 18.44±4.8 -> 10.28±1.8 | Helped |
| Add Jacobian regularization (ablation) | FM-GAN · SVHN (1000 labeled examples) | test error (%): 8.11±1.3 -> 4.74±1.2 | Helped |
| Add Jacobian regularization (ablation) | FM-GAN · CIFAR-10 (1000 labeled examples) | test error (%): 21.83±2.01 -> 20.87±1.7 | Helped |
| Add Jacobian regularization (ablation) | FM-GAN · CIFAR-10 (4000 labeled examples) | test error (%): 18.63 ± 2.32 -> 16.84±1.5 | Helped |
| Add both Jacobian regularization and GAN-estimated tangent invariance (ablation) | FM-GAN · SVHN (500 labeled examples) | test error (%): 18.44±4.8 -> 4.87 ± 1.6 | Helped |
| Add both Jacobian regularization and GAN-estimated tangent invariance (ablation) | FM-GAN · SVHN (1000 labeled examples) | test error (%): 8.11±1.3 -> 4.39 ± 1.2 | Helped |
| Add both Jacobian regularization and GAN-estimated tangent invariance (ablation) | FM-GAN · CIFAR-10 (1000 labeled examples) | test error (%): 21.83±2.01 -> 19.52±1.5 | Helped |
| Add both Jacobian regularization and GAN-estimated tangent invariance (ablation) | FM-GAN · CIFAR-10 (4000 labeled examples) | test error (%): 18.63 ± 2.32 -> 16.20±1.6 | Helped |
Closest papers by meaning
- Good Semi-supervised Learning That Requires a Bad GAN (NeurIPS 2017 · similarity 0.90)
- MarginGAN: Adversarial Training in Semi-Supervised Learning (NeurIPS 2019 · similarity 0.89)
- Improving GAN with Neighbors Embedding and Gradient Matching (AAAI 2019 · similarity 0.88)
- On Data-Augmentation and Consistency-Based Semi-Supervised Learning (ICLR 2021 · similarity 0.88)
- SMaRt: Improving GANs with Score Matching Regularity (ICML 2024 · similarity 0.86)
- Fusing Conditional Submodular GAN and Programmatic Weak Supervision (AAAI 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.