Scribd
Upload
4 views1 page

Abstract

This document presents an automated summarization system for scientific papers that combines citation graph analysis with semantic summarization. It processes input documents to extract structured information and constructs a citation network to evaluate the paper's influence within the research community. The system generates a citation-aware summary and visualizes the results through an interactive dashboard for better understanding of the paper's significance.

Uploaded by

bugzillageekman
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
4 views1 page

Abstract

This document presents an automated summarization system for scientific papers that combines citation graph analysis with semantic summarization. It processes input documents to extract structured information and constructs a citation network to evaluate the paper's influence within the research community. The system generates a citation-aware summary and visualizes the results through an interactive dashboard for better understanding of the paper's significance.

Uploaded by

bugzillageekman
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Automated Research Paper Summarization with

Citation Graph Analysis

Abstract
With the rapid growth of scientific literature, understanding not only the content
of a research paper but also its impact within the research community has
become increasingly challenging. Traditional automatic summarization
techniques primarily focus on textual content and often ignore the broader
research context in which a paper exists. To address this limitation, this project
presents an automated scientific paper summarization system that integrates
citation graph analysis with semantic summarization.
The proposed system accepts a research paper either as a PDF document or an
arXiv identifier. The input document is parsed using GROBID to extract
structured information, including the title, abstract, section-wise content,
references, and in-text citations. Citation contexts are then extracted to capture
the semantic reasons behind referencing other works. Using these citations, a
directed citation network is constructed, where papers are represented as nodes
and citation relationships as edges. Graph intelligence techniques, specifically
PageRank, are applied to identify influential papers and to assess the paper’s
position within the research ecosystem.
To reduce noise and retain only the most relevant information, semantic
retrieval techniques are employed to select key evidence from both the paper
content and citation contexts. A long-document summarization model is then
used to generate a citation aware summary that reflects not only the core
contributions of the paper but also its influence and significance within the
research community. Finally, the generated summary, citation graph, and
influence metrics are presented through an interactive dashboard, enabling users
to quickly understand both what the paper does and why it matters.

You might also like