Predicting the Past on The Getty Iris

I have a new piece out on the Getty’s Iris introducing how historians - particularly art historians - are using computational modeling and sometimes even machine learning to predict the past:

Computational simulations will never capture the full complexity of history—but then again, historians aren’t here to tell everything. We are here to distill evidence into a cogent argument. Explicit models, whether or not they use large datasets or machine learning, are a great help to historians in pursuit of this goal. Like an X-ray, a simplified model does flatten our view of the subject. But it also reveals a perspective that would otherwise have gone unseen.

See the full post at “Predicting the Past: Digital Art History, Modeling, and Machine Learning”