Professor Ng shared the results of his poll during the talk on March 24th, 2021. 10,000 was the most frequently entered number of training examples in supervised learning models trained by participants.

10,000 is a frequent machine learning training set size

1,000-10,000 training examples was the most frequent range for dataset sizes in recent Kaggle competitions.  This range supports the inference that 10,000 training examples is a common number in many real-world applications of machine learning.

University of Michigan - Ann Arbor

Andrew Ng believes that for datasets with fewer than 10,000 data points, machine learning teams can make faster progress by focusing on improving data than improving code.

MLOps: Data-centric and Model-centric approaches



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

A Chat with Andrew on MLOps: From Model-centric to Data-centric AI

According to Andrew Ng, the left side of this chart represents a traditional view in the machine learning field, which has been reinforced though events like competitions where teams focus their work on making changes to models to improve performance.  Ng posits that this emphasis has resulted in an overfocus on making adjustments to models, when improvements to data quality could lead to better results.

Andrew Ng's Comparison Chart for Data-centric and Model-centric Approaches

Andrew Ng suggests that attention to data should be proportional to its influence on AI

Data is food for AI

This chart represents changes in processes that have arisen with AI systems.  Previously, there was not a need for ML engineers or MLOps to be focused on data quality and consistency.

Learn Before

Related

Learn After