ITP Camp :: 6/21/2017
From principal components to puppyslugs
- Generative models (4:10)
- Why sampling images is hard (5:58)
- Principal component analysis (9:57)
- Eigenfaces (15:15)
- Neural networks (29:28)
- Representation learning and feature extraction (48:26)
- Autoencoders (1:01:31)
- Generative adversarial networks (1:07:31)
- BEGAN, InfoGAN, DiscoGAN, StackGAN, ArtGAN (1:17:44)
- Deep generator networks (1:21:50)
- Conditional GANs (pix2pix) (1:26:25)
- CycleGANs, horse2zebra (1:34:27)
- Skip-thought vectors and WaveNets (1:37:11)
- Class synthesis, deepdream, and puppyslugs (1:41:08)