Skip to content

srm-mic/Visual-Research-Paper

 
 

Repository files navigation

Visual Research Paper

Project for IntraMIC Hack 'September 2020


Research Papers are Boring!

Objective:

Take a pdf of a research paper and generate a comprehensive mind map as a means of visual summarization.

Motivation:

As executives in an undergraduate machine intelligence research community, we saw a need in fellow students with learning disabilities who had difficulty understanding the often tedious text of research papers. This project also proved to be extremely helpful for visual learners within the community as well. We’re hoping to make this tool open to the general public soon.

How It Works:

Components:

  1. Extract the structured text from the pdf
    Input: PDF
    1. Extract text itself
    2. Identify headings and subheadings
    Output_1: The outline of the document in two parts: 1. List of Headings 2. List of Dicts Holding Structured Plain Text
  2. Summarization of the lowest levels of content (for now: paragraph)
    Input: Output_1
    1. Using outline, extract paragraph content under each heading
    2. Use summarizer to summarize paragraphs under one heading as one.
      1. Set limit to how many words regardless of number of paragraphs
      2. If the introductory paragraph falls under a heading with no other subheading attached, include summary with heading node.
    3. Update heading line by appending summary
    4. Replace content in Output_1
    Output_2: Outline with Summarized Paragraphs
  3. Feed Output_2 into application that will generate interactive mind map
    Input: Output_2
    1. Generate the interactive mind map by creating a graph using PyVis
    Output_3: Interactive Mind Map = Visual Research Paper

To Do:

  1. Extract figures and tables to add its labels to image nodes containing figure/table
  2. Extract formulas to add as separate nodes
  3. Generate a cleaner mind map

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 77.1%
  • HTML 19.8%
  • Python 3.1%