Loss Functions
GAN Hinge Loss
Introduced by Lim et al. in Geometric GANThe GAN Hinge Loss is a hinge loss based loss function for generative adversarial networks:
$$ L_{D} = -\mathbb{E}_{\left(x, y\right)\sim{p}_{data}}\left[\min\left(0, -1 + D\left(x, y\right)\right)\right] -\mathbb{E}_{z\sim{p_{z}}, y\sim{p_{data}}}\left[\min\left(0, -1 - D\left(G\left(z\right), y\right)\right)\right] $$
$$ L_{G} = -\mathbb{E}_{z\sim{p_{z}}, y\sim{p_{data}}}D\left(G\left(z\right), y\right) $$
Source: Geometric GAN
Papers
| Paper | Code | Results | Date | Stars |
|---|
Tasks
| Task | Papers | Share |
|---|---|---|
| Image Generation | 45 | 15.46% |
| Conditional Image Generation | 17 | 5.84% |
| Reinforcement Learning | 7 | 2.41% |
| Multi-agent Reinforcement Learning | 7 | 2.41% |
| Speech Synthesis | 7 | 2.41% |
| Translation | 7 | 2.41% |
| Super-Resolution | 6 | 2.06% |
| Reinforcement Learning (RL) | 6 | 2.06% |
| Image-to-Image Translation | 6 | 2.06% |
Usage Over Time
Components
| Component | Type |
|
|---|---|---|
| 🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |
