There is increasing use of machine learning in education as it facilitates both teaching and learning. As part of artificial intelligence, machine learning is supporting teaching and learning in a number of ways.

Here we discuss three successful use cases.

1. The University of Michigan incorporates automated text analysis tool

Thanks to machine learning, there is now an algorithm to improve students’ writing.

Two professors at the University of Michigan have developed a writing-to-learn tool called M-Write, which helps students with their writing assignments.

M-Write is built on the premise that students learn better when they do writing assignments about the material that they’re studying, rather than answer multiple-choice questions about the work.

The M-Write program has now added automated text analysis (ATA) to analyze the strengths and weaknesses of writing assignments and so identify students who are struggling.

Universities have a need for an automated text analysis tool to score the writing assignments of large classes. In these cases, lecturers often resort to multiple-choice quizzes because they take less time to grade. However, it’s better to test understanding with more thorough writing assignments.

Incorporating ATA will enable lecturers to give students more involved writing assignments which will improve their learning.

To build the automated system, a software development team used data from students who previously participated in the M-Write program. The information included reviews and scores from professors and fellows. The developers used this information to write algorithms that can spot students who are struggling with the course material.

The algorithms use a variety of techniques, such as vocabulary matching or topic matching to analyze students’ texts.

2. Johns Hopkins University (USA) creates software that automates grading of tests and exams

Grading is a time-consuming aspect of teaching, especially in large courses. When lectures have to make use of a number of teaching assistants (TAs) to help with the grading, inconsistencies can creep in.

If aspects of the grading could be automated, it would eliminate inconsistencies and save time. Gradescope, developed at The University of California Berkeley, is software that automates certain aspects of grading assignments and exams.

To use Gradescope, the lecturer scans exams or tests into the platform and creates a reusable rubric to grade each question. The rubric contains the list of competencies or qualities used to assess correct answers. Making use of a rubric helps to keep the grading consistent.

Scott Smith, professor of Computer Science at Johns Hopkins University has used Gradescope. He decided to try out the software after he found out that he would not have access to TAs to help him with the grading of his Principles of Programming Languages course.

Gradescope helped him out instead. He writes, “Overall, the use of Gradescope has reduced time spent grading and improves the quality of feedback that I am able to provide students.”

3. University of California, San Diego (USA) develops an intelligent tutoring system

UC San Diego computer science professor Pavel Pevzner and colleagues have developed the first online advanced undergraduate course designed specifically as an adaptive intelligent tutoring system (ITS). It was developed for the edX platform.

The Introduction to Genomic Data Science course gives learners access to the best content in the field, and importantly, through the ITS, it provides an adaptive and personalized learning path for each online student enrolled in the MOOC.

This is achieved through quizzes and “just in time” exercises that allow for continual evaluation of each student throughout the course.

Explains Pevzner: “The goal is to represent each learner as a pathway through the tutoring system and to analyze those digital paths across thousands of students – producing an avalanche of data that can be used to continually adjust the coursework to meet the needs of all learners.”

The widespread adoption of ITS could drastically change traditional classroom teaching in the near future – no more lecture halls filled with hundreds of students exposed to the same information at the same time.

Adaptive intelligent tutoring systems will change all that.