Ram Maheshwari Logo Image
Chengyi Lux Zhang

Unique Variant

A fully configurable controlled-randomness framework for online assessment platforms.

Chengyi Lux Zhang, Mingxiao Wei, Christian Moran, James Rakanatha Yosharry Litanto, Dan Garcia

Project Overview

Priarielearn is an online learning and assessment platform originally developed by UIUC and widely used by CS courses at UC Berkeley. It is a great tool for generating a rich randomized variation of assessments, but has no support for controlled randomness. Unique variant aims to solve this problem by providing a fully configurable controlled-randomness framework that is as non-invasive to the origianl question templates as possible while being performant, easy to use, and scalable. We provide with our own adaptation of Fisher-yates shuffle algorithm that stores no additional metadata to the database.

Unique Variant opens up new opportunities ensuring students get fair assessment that is non-repetitive for each students yet maintaining the randomness across the entire class. It also facilitates the development of NextVariant, allowing instructors to high-quality Q&A the question templates comprehensively.