welcome

We now live in a world where we are spending a lot more time online due to technological advancements such as the internet and smartphones. Traditional media such as printed newspapers and DVDs has been replaced by social media and video streaming services. We conduct a lot more purchases online and participate in the service economy by using apps like Uber and Doordash.

The above highlights an increased need for timeliness in service delivery. This has spillover effects into machine learning, which is often a critical component in any kind of real-time application. There are many more use cases where generating accurate inferences in real-time has become a necessity.

This book introduces the concept of real-time machine learning, also known as online learning. Real-time machine learning uses models that are able to learn incrementally on data as soon as it arrives in a system without the need to accumulate a batch of data before training a model. These types of models are better able to adapt to changing characteristics in data to produce accurate inferences. For example, Netflix used to train its machine learning models daily to serve recommendations to its users. Netflix switched to using real-time machine learning models, which resulted in more timely and relevant recommendations and consequently kept users more engaged.