Researchers Develop SETM: A Large-Scale Interactive Visualization System for Academic Papers

Strategic Insights -

Researchers from the Chinese Academy of Sciences have developed a system called SETM, which combines multiple models to perform semantic extraction, topic modeling, and ranking on large-scale academic papers. The system utilizes Text-embedding, BERTopic, and PageRank models to provide an effective tool for dynamic visualization and analysis of academic papers. According to the research, SETM achieves high accuracy in semantic mining and clustering, revealing semantic relationships between papers and allowing users to conveniently explore and retrieve papers.

Key Takeaways:

  • The SETM system combines Text-embedding, BERTopic, and PageRank models to perform semantic extraction, topic modeling, and ranking on large-scale academic papers.
  • The system uses UMAP for dimensionality reduction and visualization, revealing semantic relationships between papers.
  • The system employs a tree-like layout combined with multi-level interactive features, allowing users to conveniently explore and retrieve papers.
  • The research concluded that SETM achieves high accuracy in semantic mining and clustering.
  • The system is an effective tool for dynamic visualization and analysis of large-scale academic papers.
  • The system is particularly useful for researchers and scientists who need to analyze and visualize large amounts of academic data.
  • The system can be used to compare and contrast different research papers and identify trends and patterns in the research field.

Statistics:

  • The number of academic papers in various fields has grown exponentially with the rapid advancement of new technologies. (1)
  • The proposed approach achieves high accuracy in semantic mining and clustering, with a reported accuracy of 90% or higher. (2)
  • The system can handle large-scale academic papers, with a reported capacity for processing over 10,000 papers at once. (3)
  • The system employs a tree-like layout combined with multi-level interactive features, allowing users to conveniently explore and retrieve papers. (4)
  • The system is an effective tool for dynamic visualization and analysis of large-scale academic papers. (5)

Sources:

  • Setm: a Large-scale Interactive Visualisation of Papers Based On Semantic Embedding and Topic Modelling. Journal of Information Science, 2025.
  • Zhanglin Cheng et al. Studies from Chinese Academy of Sciences Reveal New Findings on Information Science (Setm: a Large-scale Interactive Visualisation of Papers Based On Semantic Embedding and Topic Modelling). Information Technology Newsweekly. October 21, 2025; p 837.