Greg provides a pick of the crop of deep-learning papers from 2017 along with explanations for what makes these papers special.
The list includes:
- Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
- Wasserstein GAN
- Learning from Simulated and Unsupervised Images through Adversarial Training
- Mastering the game of Go without human knowledge
- Deep Image Prior