Manuals, Tutorials, and More

Home > Turnitin Scoring Engine > TSE Release Notes

TSE Release Notes

The Scoring Engine team is committed to keeping our users as informed with the product creation process as possible. In this light, we share these release notes with you.

 

Our release notes cover what we’re working on, things we’ve fixed, and known issues with potential workarounds.


Please read these notes before reporting a bug.

 

Release Note History:


7.08.2016

New Feature 

Cross-Validated Prediction

 

Cross-Validated Prediction files now provide the following: 

 

  • Support for custom essay IDs. These IDs are pulled from the original training data CSV. You will see these custom IDs alongside the predictions in the downloadable predictions file.

and

  • Show predictions for each trait as columns in the predictions csv, instead of being stacked on one another.

 

Likewise, non-integer IDs are also supported. 


6.30.2016

New Feature 


Create a Grader

Users can now create an automated grader to score essays based on their own self-service models. When a model achieves a level of reliability that you are comfortable with, clicking the “Create Grader” button contacts the TSE team. Users will be notified and presented with a grader to be used via API.

 

For more information on the Create a Grader feature and implementation instructions, click here


6.03.2016

Bug Fix

 

Before

Uploaded evaluations without attached datasets would not appear in the homepage list.

After

Evaluations incurring errors are now visible and can be deleted.


4.28.2016

Bug Fix

 

A bug preventing the users from scrolling within the TSE homepage has been fixed. Users are now able to expand all items and scroll throughout.


4.25.2016

Your Content. Your Rubric. Your Scoring Solution.

Turnitin Scoring Engine self-service is live! 

With the ability to accurately score an unlimited number of new compositions quickly and reliably, Scoring Engine empowers users to put their data sets to work for them. 

 

Users can: 

  • Upload data sets: Using TSE's validators, rectify training data set to ensure the best chance for accurate Grader creations. 

 

  • Evaluate data sets: With a successful upload, TSE breaks down each rubric category in relation to its grader efficacy and effectiveness. TSE will custom build a grader for your data set and provide a research-quality report.

 

  • Request grader creation: TSE will custom build a grader for individual data set. Integrate this grader with your scoring solution to get real-time score predictions on new essays.

 

Looking for more?

Product Overview

Help Center

Consultation  

You must to post a comment.
Last modified
18:24, 8 Jul 2016

Tags

This page has no custom tags.

Classifications

(not set)