Special Feature

Quizlet is betting AI, machine learning can tutor students, close homework gap

The suite of Learning Assistant product features is available to Quizlet Plus subscribers.

special report

Back to virtual school: Education embraces remote learning

Technology makes the shift possible, but challenges abound.

Read More

Quizlet is launching Quizlet Learning Assistant, a set of new product features that will combine machine learning and learning science to better tutor students.

Learning Assistant combines machine learning data with past study performance to set paths to tackles a subject. The suite of Learning Assistant product features is available to Quizlet Plus subscribers. 

"The goal is effectively guiding the student on that path to mastery leveraging learning and cognitive science and machine learning," said Quizlet CEO Matthew Glotzbach. "We are trying to approximate that capabilities a tutor can provide."

Indeed, a tutor will cost you $50 to $100 an hour relative to Quizlet Plus at $3.99 a month. 

Key components include:

  • Study paths, which are based on a short diagnostic quiz, that offer personalized guidance and challenges.
  • Smart grading, which uses natural language processing, to interpret written answers and evaluates them on meaning instead of verbiage. The use of natural language processing is to help study concepts. Students can opt-in or our of smart grading.
  • Progress, a view into study activity and areas where a student can improve. Learning Assistant Progress has been available to Quizlet Plus users.   

Glotzbach said the features are based on generalized machine learning models and then personalize and create domain specific models for language, biology and math. The study paths will also start with pretesting and assessing baseline knowledge, he said.

Ultimately, Quizlet will see a flywheel of data from its machine learning and AI tools. For instance, Learning Assistant's smart grading tool will be able to use natural language processing to assess knowledge and understanding. "We can use the answers as they come back anonymized to better train models," said Glotzbach, who noted that Quizlet spent time compressing NLP models so they can be used in real-time.

Glotzbach said Quizlet has been working on Learning Assistant's machine learning tools for the last couple of years with a dedicated team of 30 to 40 people working across engineering, data science, product and design.  

Here are two screenshots of the features with smart grading followed by the progress dashboard. 

smart-grading-example-1.png

progress-example-1-progress-dashboard.png