Automatic Quiz Generation System (AQGS)

Abstract

The thesis presents the development of an Automatic Quiz Generation System (AQGS) that uses Natural Language Processing (NLP) and Artificial Intelligence (AI) to automate quiz creation from digital documents. This system addresses the time-consuming nature of manual quiz creation and improves scalability in educational settings. The AQGS follows five key steps: extracting text from files, segmenting the text into smaller chunks, creating embeddings, storing these embeddings as vectors, and generating quiz questions using Large Language Models (LLMs). By automating this process, AQGS reduces the time and effort required for quiz creation, enabling educators to focus on more important tasks. The project also discusses technical challenges and tools like Langchain, GROBID, and Pinecone. Future improvements include code-based question generation and enhancing quiz output quality.