The difference between training and inference is similar to that of a student being taught something like algebra at school and then applying it in the real world. In school, the student is given numerous exercises; for each exercise, the student attempts the question and hands his/her answer over to the teacher, who provides feedback indicating whether it is correct or not. Initially, this feedback is likely to be skewed toward the student being wrong more often than right, but after many attempts, as the student starts building his/her understanding of the concepts, the feedback shifts towards mostly being right. At this point, the student is considered to have sufficiently learned algebra and is able to apply it to unseen problems in the real world, where he/she can be confident of the answer based on his/her exposure to the...

For those unfamiliar with the term edge computing, it simply refers to computation performed at the end, or edge, of a network as opposed to sending it to a central server for computation. Some examples of edge devices include cars, robots, Internet of Things (IoT), and, of course, smartphones. 

The motivation for performing computation at the edge, where the data resides, is that sending data across the network is expensive and time-consuming; this incurred latency and cost restrict us with what experiences we can deliver to the user. Removing these barriers opens up new applications that would otherwise not be possible. Another benefit of performing inference at the edge is data privacy; removing the need of having to transmit personal data across the network reduces the opportunities that a malicious user has for obtaining it.

Luckily, technology...

With the release of iOS 11 and Core ML, performing inference is just a matter of a few lines of code. Prior to iOS 11, inference was possible, but it required some work to take a pre-trained model and port it across using an existing framework such as Accelerate or metal performance shaders (MPSes). Accelerate and MPSes are still used under the hood by Core ML, but Core ML takes care of deciding which underlying framework your model should use (Accelerate using the CPU for memory-heavy tasks and MPSes using the GPU for compute-heavy tasks). It also takes care of abstracting a lot of the details away; this layer of abstraction is shown in the following diagram: 

There are additional layers too; iOS 11 has introduced and extended domain-specific layers that further abstract a lot of the common tasks you may use when working with image...

In Chapter 1, Introduction to Machine Learning, we saw many different types of learning algorithms and learned that ML is really a process of automatically discovering rules given a set of examples. The main components required for this process, specifically for supervised learning, include: 

• Input data points: For image classification, we would require images of the domain we want to classify, for example, animals. 
• The expected outputs for these inputs: Continuing from our previous example of image classification of animals, the expected outputs could be labels associated with each of the images, for example, cat, dog, and many more. 
• A ML algorithm: This is the algorithm used to automatically learn how to transform the input data points into a meaningful output. These derived sets of rules are what we call the...

When performing ML on the edge, you lose some of the luxuries you tend to have when running on a more powerful device (albeit this is shifting all the time). Here is a list of considerations to keep in mind:

• Model size: Previously, we walked through building a simple linear regression model. The model itself consists of two floats (bias and weight coefficients), which of course are negligible in terms of memory requirements. But, as you dive into the world of deep learning, it's common to find models hundreds of megabytes in size. For example, the VGG16 model is a 16-layer conventional neural network architecture trained on the ImageNet dataset used for image classification, available on Apple's site. It is just over 500 megabytes. Currently, Apple allows apps 2 gigabytes in size, but asking your user to download such a large file...

In this chapter, we discussed the difference between training and inference, along with the typical ML workflow and where Core ML fits in. We also saw how Core ML is not just a single framework, but rather a suite of tools that facilitate getting pretrained models into the iOS platform and making them available to your application via a familiar and simple interface. Thus, it democratizes ML and puts it into the hands of many iOS app developers.

It has been suggested that the explosion in diverse apps contributed to the success of the adoption of smartphones; if this is true, then prepare yourself for the next explosion of AI-enhanced apps. And take comfort knowing that you are in the perfect place to begin and lead this journey, where we will explore many concepts and examples related to computer vision using Core ML, including these:

• Recognizing objects...