The Knewton Blog

Our monthly newsletter features edtech and product updates, with a healthy dose of fun Knerd news.

Knewton, Simplified

Posted in Knerds on April 17, 2014 by

Most big technology companies are complex. Knewton is no exception. There’s a lot going on behind the scenes. Algorithms, data stores, model computation engines, inference infrastructures, APIs — we’ve got ’em.

But what about why we’re here?

Knewton technology exists to help everyone in the education ecosystem — including students themselves — improve learning. Knewton helps teachers differentiate instruction for every student and helps self-guided students explore more productively. Like teachers, our goal is to help prevent more students from falling behind, and give more students the chance to get ahead.

How do we do this?

  • We estimate a student’s understanding of what they’re learning, down to the concept level. When a student uses learning materials (i.e., a digital course, textbook, or app), Knewton uses lots of data to generate ongoing estimates of that student’s underlying competency in specific areas. These competencies may be defined by a particular concept (e.g., “light travels in straight lines,”) or by organizations of concepts, such as a specific state’s test requirements, the Common Core, AP prep, and any other taxonomy a content author uses to catalog subject matter.
  • We recommend what activity a student would benefit most from doing right now. Knewton suggests what a student should do right now that will best help him or her achieve a concrete goal. Recommendations open doors for students to challenge themselves and explore new areas of study. A teacher might use recommendations to build a series of “goals” and due dates for different groups of students, which Knewton uses to suggest instruction or assessment activities dynamically. Or a teacher might assign an adaptive “follow-up” module to every regular homework assignment — providing personalized reinforcement, remediation, or enrichment when a student needs it.
  • We communicate actionable information to teachers. We provide metrics that help teachers a) make sense of student learning patterns in the context of a whole class, and b) anticipate what students will need in the near future. Teachers do a lot of this already, but it becomes very difficult to do manually as students pursue increasingly varied paths. Here are a few metrics we provide, and the questions they help answer.
    • Proficiency: How proficient is Jay in the topics covered in this chapter?
    • Readiness Forecast: Given multiple factors (e.g., Jay’s current rate of learning, work completed, and current understanding), is he likely to be proficient in this entire unit in three weeks?
    • Predicted Score: What would Jay score on Friday’s quiz if he took it today?
    • Active Time: How much productive time has Jay spent working in the past week?

Today, teachers and students access Knewton through digital learning products developed by publishers and learning companies. (Soon, any individual  will be able to access these features through a free Knewton tool.)  These companies connect their products to Knewton’s technology infrastructure over the internet. The product sends anonymized, encrypted interaction data to Knewton, and Knewton sends back encrypted data for the product to use.

So that’s what we do. But our technology is just part of a much larger ecosystem (instructors, user interface, instructional design, content, pedagogy, etc). In the end, teachers are the ones with the experience, emotional intelligence, and cultural understanding to know what will work for their students. We all share the same goal: to help every student maximize their own educational potential.