Quizzer an Assessment Generation Portal using NLP

Project Title

Quizzer an Assessment Generation Portal using NLP

Project Area of Specialization Artificial IntelligenceProject Summary

Quizzer in an assessment generation portal based on the principles of Natural Language Processing (NLP) which is a sub-division of Artificial Intelligence (AI). The main idea behind Quizzer is to create a web portal that takes text (or PDF files) from users and converts it into Fill in the Blanks, True/False statements, and multiple-choice questions (MCQs). Quizzer is smart enough to detect what kind of text it's given, and it then creates the aforementioned assessment problems from it.

Project Objectives Quizzer's Objectives:

• To enable students in performing self-assessments, especially during this era of remote learning.
• Assisting instructors in test creation. It saves time by allowing facilitators to create assessment problems on the go.

  Project Implementation Method

The main idea is to create a web portal that allows access to three types of users: students, teachers, and guest users.

Guest User:

• Can generate assessments (MCQs, Blanks, and T/F statements) from any sort of (non-technical, and entry-level technical) text for self-assessment.

Students:

• Can generate assessments (MCQs, Blanks, and T/F statements) from any sort of (non-technical, and entry-level technical) text for self-assessment.
• Can attempt any quizzes created by teachers using a unique quiz key/ID.

Teachers: 

• Can generate assessments (MCQs, Blanks, and T/F statements) from any sort of (non-technical, and entry-level technical) text for self-assessment.
• Can create and grade quizzes.

How do we plan on making this happen? Let's explore the necessary categories: 

MCQs: A Sense2Vec model will be used to generate distractors. Distractors are the wrong choices that we see in MCQs. Simply, a keyword will be extracted from the given text using Python's Keyword Extraction (PKE) tool, and the aforementioned Sense2Vec model will generate accurately misleading distractors for it.

True/False Statements: Constituency parsing is the main technique that’ll be used, we’ll be using it to split compound sentences into simpler ones. An OpenAI-GPT 2 model will be used for text generation; this model will add further text to the initially extracted simple sentence.

Fill in the Blanks: Text ranking algorithms like TextRank and Multipartite Rank will be used to rank keywords from the given text,/these keywords will then be extracted using PKE or Flashtext tool, depending on the system's need.

Web Application: The web application will be coded in Python using the Django Framework. CSS and HTML are used for beautification.

It should be noted that all this text is processed once the user enters it, the text goes through a summarizer.

Benefits of the Project

The project is perfect for self-assessment, and assessment in general. Just the idea of not having to rely on external resources like guess papers is reason enough to consider this a success. From an instructor's perspective though, this will save them valuable hours that they can invest in something else.

Technical Details of Final Deliverable

The final deliverable will be a deployed web portal that can host as many users as need be. All three types of users must be able to access it with minimum delays and downtime. 

Final Deliverable of the Project Software SystemCore Industry EducationOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Quality EducationRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	55000
System upgrades for running the hosted website alongside the QG Model	Equipment	1	50000	50000
Project Report and Documentation	Miscellaneous	1	5000	5000