According to Professor Ng, the best reason for machine learning teams to shift from being more model-centric to more data-centric is:

University of Michigan - Ann Arbor

Andrew Ng believes that a key way to improve data consistency is to direct a good labeler to use a strategy employed by another good labeler in cases where there are multiple correct ways to categorize data.  Then it is important to continue to adjust the rules for the algorithms until consistent results are produced.

Data-centric approach: systematically fixing correct labels, Part 2

DeepLearningAI. (2021, March 24). A Chat with Andrew on MLOps: From Model-centric to Data-centric AI. YouTube. https://www.youtube.com/watch?v=06-AZXmwHjo

Learn Before

Related