Summarizing the book

We started with supervised learning approaches and focused on how to create classification models. In particular, we saw how it's possible to do the following:

• Use a perceptron for a linearly separable problem (Chapter 2, Neural Network Fundamentals)
• Use feedforward neural networks (FFNNs) for non-linearly separable tasks (Chapter 3, Convolutional Neural Networks for Image Processing)
• Use embeddings to extract useful information from text (Chapter 4, Exploiting Text Embedding)
• Use Convolutional Neural Networks (CNNs) for tasks whose inputs have a spatial relationship (Chapter 5, Working with RNNs)
• Use pre-trained (Neural Network) (NN) as a feature extractor (Chapter 6, Reusing Neural Networks with Transfer Learning)
• Use generative models to reproduce the creativity process (Chapter 7, Working with Generative Algorithms, Chapter 8, Implementing Autoencoders...