1. Visualizing CNN

1-1. What is CNN visualization

• CNN is a black box
  • What is inside CNNs?
  • Why do they perform so well?
  • How do we improve them?

1-2. Vanilla example: filter visualization

1-3. How to visualize neural network

2. Analysis of model behaviors

2-1. Embedding feature analysis

1) Nearest neighbors in a feature space

2) Dimensionality reduction

• t-distributed stochastic neighbor embedding (t-SNE)

2-2. Activation investigation