Deep Learning Life Cycle
In this chapter, we will explore the intricacies of the deep learning life cycle. Sharing similar characteristics to the machine learning life cycle, the deep learning life cycle is a framework as much as it is a methodology that will allow a deep learning project idea to be insanely successful or to be completely scrapped when it is appropriate. We will grasp the reasons why the process is cyclical and understand some of the life cycle’s initial processes on a deeper level. Additionally, we will go through some high-level sneak peeks of the later processes of the life cycle that will be explored at a deeper level in future chapters.
Comprehensively, this chapter will help you do the following:
- Understand the similarities and differences between the deep learning life cycle and its machine learning life cycle counterpart
- Understand where domain knowledge fits in a deep learning project
- Understand the few key steps in planning a deep learning project to make sure it can tangibly create real-world value
- Grasp some deep learning model development details at a high level
- Grasp the importance of model interpretation and the variety of deep learning interpretation techniques at a high level
- Explore high-level concepts of model deployments and their governance
- Learn to choose the necessary tools to carry out the processes in the deep learning life cycle
We’ll cover this material in the following sections:
- Machine learning life cycle
- The construction strategy of a deep learning life cycle
- The data preparation stage
- Deep learning model development
- Delivering model insights
- Managing risks
