1.Introduction

2. Related Work

3. Jacobians of Neural Networks

3.1. Special case: ReLU and MaxPool

3.2. Invariance to weight and architecture specification

4. Distillation