Summary 

In this chapter, we introduced object recognition, the 101 project for ML with Core ML. We spent some time introducing CNNs or ConvNets, a category of neural networks well suited for extracting patterns from images. We discussed how they build increasing levels of abstraction with each convolutional layer. We then proceeded to make use of our newfound knowledge by implementing the functionality that allowed our application to recognize the physical world through its camera. We saw firsthand that the majority of the work wasn't performing inference but rather implementing the functionality to facilitate and make use of it. This is the take-away; intelligence by itself is not useful. What we are interested in exploring in this book is the application of trained models to deliver intuitive and intelligent experiences. For instance, this example can...