Chapter 7: Machine Learning Workloads at the Edge

The growth of edge computing is not only driven by advancements in computationally efficient hardware devices but also by the advent of different software technologies that were only available on the cloud (or on-premises infrastructure) a decade back. For example, think of smartphones, smartwatches, or personal assistants such as Amazon Alexa that bring a mix of powerful hardware and software capabilities to consumers. Capabilities such as unlocking your phone or garage doors using facial recognition, having a conversation with Alexa using natural language, or riding an autonomous car have become the new normal. Thus, a need for cyber-physical systems to build intelligence throughout their lifetime based on continuous learning from their surroundings has become key for various workloads in today's world.

It's important to realize that most of the top technology companies (such as Apple, Amazon, Google, and Meta, formerly...