If you need questions, we have all the answers

A fast generator with a user friendly interface that uses various NLP models to generate high quality questions and answers.

Click for the video! Click for the poster!

Problem

Testing knowledge is a traditional requirement. These can be standardized exams or subject-specific tests. Generating new and varied questions takes effort. Also, knowledge of the subject is required and a certain amount of time is needed.

Solution

We aim to support online learning as well as face-to-face learning techniques. While serving educational and training purposes, it will be waste of time and effort to do this in an non-automated way. We set out with the idea of the automating the question generation. Request uses natural language processing to take to desired material as input and automatically generate questions and answers.

Our Features

Generating questions and answers regarding the input text.

Ability to fetch large sets of questions, answers

Three types of profiles with access to different Features

Different question types

Ability to upload own datasets

Finetuning Capability with Datasets

Collecting Feedback to Rank Generated Questions and Answers

Ability to save and delete Question-Answer pairs and Question-Answer pairs in downloadable format

Powered By

The largest hub of ready-to-use datasets for ML models with fast, easy-to-use and efficient data manipulation tools.

Hugging face is compatible with T5 architecture used in Request. Hugging face, which is an advanced and easy-to-use library, allows to apply published models.

MongoDB empowers innovators to create, transform, and disrupt industries by unleashing the power of software and data. ReQuest keeps the necessary information for NLP models and user accounts in MongoDB. It is a database of simple setup, cost-effective and flexible document schemas.

Design simple views for each state in your application, and React will efficiently update and render just the right components when your data changes.

ReQuest frontend is based on React. It simplifies the overall process of writing components, enabling faster rendering. It guarantees stable code and is SEO friendly.

Docker makes development efficient and predictable. Docker is a tool that makes the deployment process very easy for ReQuest. Docker was used to isolate nested processes.

Django makes it easier to build better web apps more quickly and with less code. Django implemented in Python is used for ReQuest backend development. Django offers fast processing, fast development. There is a great set of libraries.

Our Team

Meet Our Team

Supervised by : Pınar Karagöz

karagoz@ceng.metu.edu.tr

Ecem Gül

ecem.gul@metu.edu.tr

Adnan Harun Doğan

adnan.dogan@metu.edu.tr

Kerem Serttaş

serttas.kerem@metu.edu.tr

Umur Görkem Karaduman

gorkem.karaduman@metu.edu.tr

Zeynep Dokuz

zeynep.dokuz@metu.edu.tr

REQUEST