Short Quiz Generation Using A Multi Level Transfer Model
- Mentors
- Bruno, MiHarsh, Divyanshu_Singh
- Organization
- AOSSIE
- Technologies
- python, javascript, html, css, pytorch, HuggingFace, python-server
- Topics
- machine learning, natural language processing, software development, edtech
Online learning has taken the front seat in the post pandemic age. With the advent of sophisticated AI architectures like the Transformers, it is but natural that AI would find its way into education. Learning online via YouTube or MOOC platforms is often done as a method of self learning. The biggest obstacle faced by a student in self learning is the lack of attention span. An online tool that can generate short quizzes on input educational content can be of great use to teachers and students alike as it can help retain important information, frame questions and quickly revise large chunks of content.
EduAid is one such project where I propose the usage of open source and lightweight Large Language Models for quiz generation. The language modelling capabilities of LLMs can be leveraged for achieving this task. Finally, we develop a chrome extension that allows end users to interact with the pipeline and generate quizzes based on input text.