Abstract

Identifying the most relevant text information from online texts has become a standard task now that data is easily accessible. Conse- quently, summary generation has gained significant importance by reducing the amount of text a person has to read to get meaningful information from the text. The underlying field for this process Natural Language Processing (NLP). NLP enables algorithms to understand and rewrite a text based on the original information. In this proposal, I will explore how a Convolutional Neural Network (CNN) can create a new approach for visualizing the critical aspects of a text. Instead of generating a shorter text, the output will be a version with the essential elements highlighted.

Poster

Video

Paper

Architecture Diagram

Gitlab: https://code.cs.earlham.edu/jejunco20/capstone