Scholarlink

This paper presents an AI-powered platform designed to automate the generation of diverse assessment materials from PDF documents and raw text inputs, addressing the time-intensive challenge of manual quiz creation in education. Leveraging a fine-tuned T5-base transformer model and advanced Natural Language Processing (NLP) techniques, the platform generates Multiple-Choice Questions (MCQs), Short Answer questions, Fill-in-the-Blanks, Matching tasks, and Summaries through a user-friendly Flask-based web interface. Evaluated on the SQuAD v1.1 dataset, the platform achieved ROUGE-1 and ROUGE-L scores of 0.5247 and 0.4844, respectively, indicating strong semantic content capture. Compared to existing tools like Quizzzy and LEAF, it offers broader question type support and greater input flexibility, demonstrating significant potential to reduce educator workload and enhance learning environments. Limitations include challenges in PDF text extraction and distractor quality, with future improvements planned.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI-Powered Platform for Automated Quiz and Question Generation “QuiGen”

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ahmad Yasser Saighi,
  • Ayoub Aissaoui,
  • Akram Bettahar,
  • Larbi Guezouli,
  • Khalil Bourek

摘要

This paper presents an AI-powered platform designed to automate the generation of diverse assessment materials from PDF documents and raw text inputs, addressing the time-intensive challenge of manual quiz creation in education. Leveraging a fine-tuned T5-base transformer model and advanced Natural Language Processing (NLP) techniques, the platform generates Multiple-Choice Questions (MCQs), Short Answer questions, Fill-in-the-Blanks, Matching tasks, and Summaries through a user-friendly Flask-based web interface. Evaluated on the SQuAD v1.1 dataset, the platform achieved ROUGE-1 and ROUGE-L scores of 0.5247 and 0.4844, respectively, indicating strong semantic content capture. Compared to existing tools like Quizzzy and LEAF, it offers broader question type support and greater input flexibility, demonstrating significant potential to reduce educator workload and enhance learning environments. Limitations include challenges in PDF text extraction and distractor quality, with future improvements planned.