What is Serving to Monitor in Machine Learning

Serving to Monitor is defined as ensuring your production and training data are handled the same way, and that your production and training data are not drastically different. It is essential to measure and monitor your models in production, in the real world. 

This step is extremely important as it will enable quick detection and mitigation of issues (such as prediction drift to Bias & Data Integrity Issues). There are a number of tools you can use to monitor your models, including Aporia’s customizable ML monitoring solution.