Inspiration
During this covid-19 pandemic situation we all switched to Online mode of education and our examination mode also shifted from subjective type to objective type. In order to make our learning more fruitful one requires to do assessment, learning is not complete without assessment. As “Modern problem requires modern solution” so we are making an automated MCQ generator which will be beneficial not only for students but also teachers
What it does
An automatic generation of multiple choice questions on any user-defined domain is proposed. It first takes the content from user in text format and then this content is summarised. The sentences in the summary are transformed into stem for the MCQs. The distractors are generated using similarity metrics such as hypernyms and hyponyms. The system also generates analogy questions to test the verbal ability of the students.
How we built it
We used Google's open sourced pre-trained Natural Language Processing based Bert algorithm to summarize the text and for getting the word sense we used wordnet API in the backend. We used Django Framework for our web application.
Challenges we ran into
Connecting frontend and backend, tuning the parameters for obtaining accurate results, handling errors.
Accomplishments that we're proud of
successfully build a running website which will help to generate questions.
What we learned
Learned about Django Framework, learned about NLP and BERT , building various components of webpage and team work.
What's next for Automated Question Generation and Summarizer
Support for other languages for generating questions using M-BERT(Multi Lingual), adding new functionalities such as True or False and WH question generation.
Log in or sign up for Devpost to join the conversation.