Journalists: Learn machine learning with free hands-on videos

Try your hand at sorting images, searching videos, and analyzing text with AI.

Journalists can now learn how to use machine learning for their investigations with several videos and online coding notebooks — for free.

Late last year I taught the online class “Hands-on Machine Learning Solutions for Journalists” through the Knight Center for Journalism in the Americas. One of the lovely parts about working with the Knight Center is that once the class is over, I’m free to post the videos online.

The 15 lessons range from 4 1/2 minutes to about 15 mins, and start with an introduction about when AI might help you. They then walk you through a series of example projects tailored for journalists — such as detecting objects in images, sorting documents into piles, and extracting people’s names from a trove of text.

The projects are based on a set of online notebooks we’ve created to work with Google’s free Colaboratory service. You won’t get the quizzes, collaborative forums, or certificate of completion available to the original students, but you can still learn a lot on your own.

The videos and notebooks draw on what we’ve learned at Quartz over the last year in the Quartz AI Studio, which was funded by a grant from the Knight Foundation.

Many of the lessons use the fast.ai machine-learning library for Python, which was built to make ML easier for people not trained in math or computers. And these videos could be your journalist-oriented primer for the excellent — and free — fast.ai course Practical Deep Learning for Coders, which has helped me immensely.

If you have thoughts or questions, or find the videos useful in your own work, tweet at me at @jkeefe.