- Introduction and overview
- Regularization
- Optimization
- Initialization
- Convolutional Networks and Natural Language techniques
- [TODO RNNs]: General LSTM overview, LSTM backprop derivation, LSTM dropout (Bayer and Osendorfer, 2014; Pascanu et al., 2014a), LSTM intro blog post, Karpathy LSTM. Also RBMs, though they are weaker than RNNs (Hinton Coursera lecture 2) notes1. Attention in goodfellow 12.4.5.1, here, too. d
- [TODO VAEs]: notes1, notes2, notes3, notes3