Machine Learning Workflow

 

Machine learning workflow is the visualization of common machine learning, deep learning and AI tasks. Note this flow chart is cyclical. Many iterations may be needed to improve the model. For example, the clean data, train and evaluate phase is often repeated many times. 

Please note: these high level workflow illustrations abstracted away detailed tasks like : data preparation, data preprocessing, data modeling. 

There are many variations and flavors of machine learning model workflow. For example, Ask Question —> Gather Data —> Create Features —> Normalize/Scale —> Sample Data —> Training Data —> Create Model —> Evaluate Model  —> Testing Data —> Deploy Model. 

This is not the only way to visualize the workflow. Many variations of this chart exist and can capture nuances of different ML systems. For example, the Google ML workflow looks like this

Understand the most important data structure of deep learning and machine learning.  Click here. 

Just like you can prototype a startup tech product, you can also prototype a data product.