Thoughts on data science, education, and economics. var _gaq = _gaq || []; _gaq.push([’_setAccount’, ‘UA-30373860-1’]); _gaq.push([’_trackPageview’]); (function() { var ga = document.createElement('script’); ga.type = 'text/javascript’; ga.async = true; ga.src = ('https:’ == document.location.protocol ? 'https://ssl’ : 'http://www’) + ’.google-analytics.com/ga.js’; var s = document.getElementsByTagName('script’)[0]; s.parentNode.insertBefore(ga, s); })();
With millions of problems attempted per day, Khan Academy’s interactive math exercises are an important and popular feature of the site. (Math practice. Popular. Really!) For over a year, we’ve used a statistical model of each student’s accuracy on each exercise as the basis for awarding proficiency . When a user reaches a high enough accuracy on an exercise, we award the proficiency badge, which…
When building supervised learning models, I often come across situations where there is a large class of features I’d like to try adding, but the corresponding increase in model parameters seems daunting. For example, in my work at Khan Academy, many user models might benefit from knowing whether each our 4,000 videos has been watched by the user or not. I usually end up trying to engineer a…
While I certainly am nowhere close to giving up, in the limited experience I have comparing cognitively-based models versus, let's say, more purely empirical models, the purely empirical models have performed much better.
Online education is hot field right now. That means a lot of programmers and data scientists are interested in learning about education and instruction; a lot of teachers are rethinking the potential uses of technology and analytics. What may be less apparent or accessible to the newly interested, though, is the fantastic body of research available to them through the field of learning science(s)…
Bayesian networks (and probabilistic graphical models more generally) are cool. We computer geeks can love ‘em because we’re used to thinking of big problems modularly and using data structures. But better than being cool, they’re useful . Especially if you have the kind of problem that involves hundreds or thousands of interrelated variables, any one of which you might want to predict based on…