Scholarlink

With the rapid increase in the volume of scientific literature, researchers face challenges in keeping up with the latest advancements while summarizing the documents. Scientific document text summarization offers a solution by providing concise and informative summaries that highlight the key contributions from original texts. This study introduces a novel method leveraging deep learning, specifically the sBERT model to summarize scientific documents. The proposed approach treats the extractive summarization as a classification problem using a dual BERT model setup. The methodology is evaluated using data set from CL-SciSumm. Results indicate that our approach significantly outperforms the existing methods in terms of ROUGE scores, demonstrating its effectiveness in generating accurate summaries of scientific literature.

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

Citation Based Scientific Document Summarization Using Deep Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Divya Jyoti,
  • Jyoti Srivastava,
  • Dharmendra Prasad Mahato

摘要

With the rapid increase in the volume of scientific literature, researchers face challenges in keeping up with the latest advancements while summarizing the documents. Scientific document text summarization offers a solution by providing concise and informative summaries that highlight the key contributions from original texts. This study introduces a novel method leveraging deep learning, specifically the sBERT model to summarize scientific documents. The proposed approach treats the extractive summarization as a classification problem using a dual BERT model setup. The methodology is evaluated using data set from CL-SciSumm. Results indicate that our approach significantly outperforms the existing methods in terms of ROUGE scores, demonstrating its effectiveness in generating accurate summaries of scientific literature.