The Economist has commented on the irony that machine learning helps school teachers relax after work by choosing what movie to watch, but helps none in determining how to assist their students. In my fellowship this summer, I tried to change that. At DSSG, I worked with the Tulsa Public Schools to do identify which students are at risk of being made to repeat 3rd grade. Using machine learning techniques we were able to predict 95% of the second grade students that would require intervention before the Reading Sufficiency Act destined them to do the year over.
At Data Fest 2016 I gave a fuller yet concise explanation, watch the video below.
If you work in early education related research with ideas about how machine learning can fit in I'd love to talk with you.