PowerGrader is an AI-assisted grading tool that integrates with the Canvas Learning Management System (LMS), also referred to as a Virtual Learning Environment (VLE).  Instructors create assignments and rubrics in the LMS.  Students submit their completed assignment to the LMS.  When instructors launch PowerGrader from the LMS, they see a list of submissions.  PowerGrader automatically grades each submission based on the rubric and provides relevant feedback to help students know why they received a grade and how they can improve.  Instructors can override grades and feedback, and PowerGraders adapts to the instructor’s grading style (more harsh or lenient) for that assignment.  When the instructor has finished reviewing the grade and feedback, they can publish the final result – which transfers the grade and feedback back to the LMS where it is visible to students. 

PowerGrader saves instructors up to 75% of the time spent grading, freeing them up to focus on lesson plans, teaching, and helping students.  It works for all text-based assignments – everything from computer science programs to essays.  Currently, PowerGrader is integrated with Instructure’s Canvas LMS.  We plan to enable integration with additional learning management systems, such as D2L’s Brightspace, in the near future.

 

Integrating PowerGrader with your LMS

Contact your Apporto Customer Success representative to integrate PowerGrader with your LMS.  This usually requires a quick call between our engineer and your LMS Admin to exchange two developer keys, an LTI key and an API key to load data like the rubric and submissions and to push grades back to the LMS.

Launch PowerGrader

Log in to your LMS and navigate to the desired course.

Click on PowerGrader in the left nav menu.

PowerGrader will open in a separate browser tab.

View the Assignments List

When PowerGrader loads, you will see a list of Assignments.  All assignments compatible with PowerGrader will be visible.  This includes all text-based assignments (e.g. code, text files, etc.).  It does not include spreadsheets, video, or images yet.  However, a PDF or text document that includes an image is ok though.  Quizzes and discussions are filtered out.

Locate the assignment you would like to grade.

Click View to view submissions for that assignment.

View the Submissions List

This Submissions screen has three sections:

1. Submissions list
2. Assignment instructions
3. Assignment rubric

The Submissions list is a list of students who have submitted the completed assignment to the LMS.  The “Reviewed and Published” status indicates whether the instructor has reviewed the submission and published the final grade back to the LMS.

The assignment Instructions and Rubric mirror those in the LMS and are shown for reference.

Click the Start Reviewing button in the upper right corner to begin reviewing submissions.

Review Submissions

The first unreviewed submission opens. 

The student’s homework file is shown on the left.  Submission scores and feedback are shown on the right along with the rubric.

Note: In education, feedback may also be referred to as comments or assessments.

A counter is shown so you can track progress (e.g. 10 or 12 submissions).

You may expand the right side (and hide the student’s homework file) by clicking the expand button in the upper right corner.  

If you ever need to reference the assignment instructions, click Open Instructions.

A sparkle symbol indicates the score was generated by AI – whereas an instructor symbol indicates it was given by the instructor.

         

AI-generated scores and feedback are provided along with the rubric.

You may edit the score (via the Custom Score text box) as well as the feedback (via the pencil edit button), and PowerGrader will adaptively learn and proactively update the scores and feedback for all other unpublished submissions for this assignment.  In other words, PowerGrader will learn your grading style – more lenient or harsh – and adapt.

In the feedback, you may see underlined text.  Click on the underlined text to highlight the relevant section of the submission – to see what prompted PowerGrader to provide the feedback. 

Note: The color of the underline does not have any particular meaning (e.g. red is not bad, green is not good).  The color is randomized.

.   

Additionally, you may click on the highlighted section to see more details about why PowerGrader provided the corresponding feedback. 

If you have made edits but have not finished reviewing a submission, click Save to save your changes.

When you have finished reviewing a submission, click Publish to publish the final score back to the LMS.

Note: Multiple instructors (e.g. professors and TAs) can be on the PowerGrader account, but only one may grade a submission and the most recent grade is stored.

Viewing Published Grades in the LMS

Login to the LMS (e.g. Canvas) and navigate to the course gradebook.

Locate the relevant student and assignment.  Then, click on the submission to view the published grade (e.g. 7/10).

Data Privacy and Security

PowerGrader does not use any of the student’s or instructor’s data to train our AI model.

PowerGrader uses an enterprise version of Open AI’s ChatGPT 4.0, and Open AI is not permitted to store our data or use it for training.

Open AI is SOC2 compliant.  Their privacy statement can be viewed here: https://openai.com/enterprise-privacy/