Scholarlink

In educational institutes worldwide, the process of formulating questions for assessments and educational materials poses a recurring challenge. Traditional methods often involve manual question creation or keyword-based searches, leading to time-consuming and often inefficient processes, especially when dealing with extensive PDF documents. Addressing this prevalent issue, the research introduces a comprehensive approach to question–answer generation (QAG) from PDF documents. The system uses the most recent Natural Language Processing (NLP) models to extract text from PDFs using pdfplumber, create document summaries using BERT, extract keywords with RAKE, and create questions using a retrained T5 model. DistilBERT is then used to generate the answers. The RQUGE metric is used to assess the system’s performance. The outcomes indicate how well the suggested method generates pertinent queries and precise responses, highlighting its potential for use in information retrieval and document comprehension.

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

Enhancing Question Answer Generation from PDFs: A Fusion of BERT, RAKE, T5 and DistilBERT with RQUGE Evaluation

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • A. C. Akshitha Mary,
  • Prachi Acharya,
  • R. Rakshinee,
  • Stefani Jeyaseelan,
  • Thanvitha Nandakumar,
  • V. Sakthivel,
  • P. Prakash

摘要

In educational institutes worldwide, the process of formulating questions for assessments and educational materials poses a recurring challenge. Traditional methods often involve manual question creation or keyword-based searches, leading to time-consuming and often inefficient processes, especially when dealing with extensive PDF documents. Addressing this prevalent issue, the research introduces a comprehensive approach to question–answer generation (QAG) from PDF documents. The system uses the most recent Natural Language Processing (NLP) models to extract text from PDFs using pdfplumber, create document summaries using BERT, extract keywords with RAKE, and create questions using a retrained T5 model. DistilBERT is then used to generate the answers. The RQUGE metric is used to assess the system’s performance. The outcomes indicate how well the suggested method generates pertinent queries and precise responses, highlighting its potential for use in information retrieval and document comprehension.