Research map / Probabilistic and scientific modeling / Variational and generative sampling
Adversarially Learned 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 |
|---|---|---|---|
| Use ALI without feature matching instead of GAN with feature matching | Semi-supervised SVHN · SVHN test set | misclassification rate: 8.11 ± 1.3 -> 7.42 ± 0.65 | Helped |
| Use ALI without feature matching instead of GAN with feature matching | Semi-supervised CIFAR10 · CIFAR10, 1000 labeled examples | misclassification rate: 21.83 ± 2.01 -> 19.98 ± 0.89 | Helped |
| Use ALI without feature matching instead of GAN with feature matching | Semi-supervised CIFAR10 · CIFAR10, 2000 labeled examples | misclassification rate: 19.61 ± 2.09 -> 19.09 ± 0.44 | Helped |
| Use ALI without feature matching instead of GAN with feature matching | Semi-supervised CIFAR10 · CIFAR10, 4000 labeled examples | misclassification rate: 18.63 ± 2.32 -> 17.99 ± 1.62 | Helped |
| Use ALI without feature matching instead of GAN with feature matching | Semi-supervised CIFAR10 · CIFAR10, 8000 labeled examples | misclassification rate: 17.72 ± 1.82 -> 17.05 ± 1.49 | Helped |
| Jointly train inference and generation instead of training an encoder after GAN training (ablation) | ALI · 2D Gaussian mixture toy dataset | modes covered: 10.4 ± 9.2 -> 13.4 ± 5.8 | Helped |
| Learn inference jointly with the decoder instead of using a learned inverse mapping from GAN samples (ablation) | ALI · 2D Gaussian mixture toy dataset | latent prior coverage and organization | Helped |
| Learn inference jointly with the decoder instead of learning inference post hoc (ablation) | ALI · 2D Gaussian mixture toy dataset | mode coverage | Helped |
| Use a VAE objective instead of ALI adversarial training (ablation) | 2D Gaussian mixture toy dataset · 2D Gaussian mixture toy dataset | reconstruction quality | Mixed |
Closest papers by meaning
- Generalized Adversarially Learned Inference (AAAI 2021 · similarity 0.91)
- It Takes (Only) Two: Adversarial Generator-Encoder Networks (AAAI 2018 · similarity 0.90)
- Structured Generative Adversarial Networks (NeurIPS 2017 · similarity 0.88)
- GibbsNet: Iterative Adversarial Inference for Deep Graphical Models (NeurIPS 2017 · similarity 0.87)
- Adversarial Feature Learning (ICLR 2017 · similarity 0.87)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 2024 · similarity 0.85)
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.