Part 3. One step beyond

In part 1 of this book, you got a basic understanding of what search engines and deep neural networks are, how they work, and how they can work together to create smarter search engines. Part 2 dove into the technical details of major deep neural network applications for search engines, mostly using recurrent neural networks and word/document embeddings to give users more relevant results. In this part of the book, we’ll tackle more-advanced topics and challenges by extending the applications of neural networks to two new areas: searching text in multiple languages using machine translation (chapter 7), and searching for images using convolutional neural networks (chapter 8). Finally, in chapter 9, we’ll look at the thing that makes the biggest difference in production scenarios: performance, whether plain speed when training and predicting, or accuracy of results. You’ll see an example of how to tune a neural network model to reach good accuracy in a reasonable training time. In addition, we’ll look at how to deal with continuous streams of data for neural search.