Understanding the ML application life cycle
The life cycle to create, maintain, and update an ML application can be visualized in a few different ways. The following diagram shows a high-level view of this ML life cycle for a typical ML project being deployed on edge devices. Take a minute to look through the listed steps before we explain each section:
Figure 6.1 – Typical ML life cycle for embedded projects
The steps at the top all pertain to data. Gathering it and preparing it for use in ML algorithms is a non-trivial task and is often the source of competitive advantage for companies (as opposed to the ML algorithms themselves). Commonly, there is a dedicated resource—or resources—for preparing data at companies, referred to here as data scientists. Once the data is prepared, the data is used to create and train an ML model. This is commonly performed by an embedded software developer and—potentially—an ML developer that...
