In other words, Data-Centric AI focuses on updating the data to solve a problem versus changing the algorithm or code. That’s a complete reversal of how we’ve thought about AI up until now. 

Over the last decade, researchers focused on code and algorithms first and foremost. They’d import the data once and generally leave it fixed. If there were problems with noisy data or bad labels they’d usually work to overcome them in the code. Data-Centric AI flips that on its head and says we should fix the data itself. Clean up the noise. Augment the dataset to deal with it. Re-label, so it’s more consistent.

The simple chart here neatly illustrates the difference between model-centric and data-centric AI.