Scholarlink

Large language models (LLMs) have revolutionized natural language processing (NLP) tasks, yet they struggle with handling long text inputs due to token limitations. This paper introduces a novel method for creating AI-driven literature surveys by integrating clustering techniques with a modified TF-IDF formula (c-TF-IDF). Our approach organizes documents into clusters, enabling the production of comprehensive and concise literature surveys. This method not only overcomes token limitations but also improves the relevance and coverage of the generated surveys. Experimental results indicate that our approach significantly outperforms traditional LLM methods in both coverage and informativeness.

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

Cluster-Based Effective Generation of AI-Driven Literature Surveys

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Zongyue Li,
  • Xiaofei Lu,
  • Jing Chen,
  • Haishan Wang,
  • Xu Wang,
  • Qinghui Shi,
  • Dejun Xue,
  • Yanhong Bi,
  • Zixuan Huang

摘要

Large language models (LLMs) have revolutionized natural language processing (NLP) tasks, yet they struggle with handling long text inputs due to token limitations. This paper introduces a novel method for creating AI-driven literature surveys by integrating clustering techniques with a modified TF-IDF formula (c-TF-IDF). Our approach organizes documents into clusters, enabling the production of comprehensive and concise literature surveys. This method not only overcomes token limitations but also improves the relevance and coverage of the generated surveys. Experimental results indicate that our approach significantly outperforms traditional LLM methods in both coverage and informativeness.