Part 5. Advanced/optional topics

 

The first four parts of this book applied to every type of AI assistant. This final part applies only to some types. However, even if you don’t need these topics directly, understanding them will make you a stronger AI assistant developer.

Chapter 11 demonstrates building a new classifier from scratch in Python code. Most AI platforms come with good classifiers that only require training, no coding required. Nonetheless, understanding how classifiers are built will give you a better appreciation for how and why training and testing work the way they do. This chapter builds a naive “bag of words” classifier to demonstrate the general concepts. I mastered classification concepts by building my own classifier. I hope this chapter helps you as well.

Chapter 12 teaches you how to train and test speech recognition for voice assistants. Voice assistants are a popular type of conversational AI as companies integrate automated solutions into their telephony systems. Voice assistants often require custom speech training. This chapter shows you how to gather speech data, as well as three different methodologies for training a custom speech recognition model.