Automated Information Search by Semantic Meaning of Text

Description:

Inventor: Jee Woong Park

He has earned a ​Ph.D. in Civil and Environmental Engineering from Georgia Institute of Technology, M.S. in Civil and Environmental Engineering from Stanford University and B.S. in Civil and Environmental Engineering from the University of Hawaii at Manoa. His work experience includes CEEC Undergraduate Coordinator at the University of Nevada, Las Vegas as of 2023, Associate Professor at the University of Nevada, Las Vegas (2023, Assistant Professor at the University of Nevada, Las Vegas (2017 – 2023) and Research Assistant at Georgia Institute of Technology (2013 – 2017)

Invention Disclosure:

An AI-based system to complete this task of reading and analyzing narratives automatically without requiring manual reading of extensive text information. The system does not require any manual reading by humans, but it automatically identifies records/information relevant to the information dictated by the user’s search criteria. By doing so, it significantly facilitates the search for information in large databases of crash narratives. The search criteria describe the crash circumstances of interest. To help analysts quickly identify relevant information in the narratives, the approach highlights the phrases more relevant to the search criteria. 

Benefits: 

  • AI-based semantic similarity module that matches the search criteria and the narratives based on their semantic contents instead of the words they contain. 
  • Enable the matching of information between search criteria and text information (narratives) based on the semantic ideas expressed in the text contents (semantics), which performs considerably better than conventional techniques in practice using the matching of individual words used to describe the ideas, 

Market Opportunity

Reading massive datasets of narratives is substantially challenging, and requires considerable time and effort to manually extract information with the naked eye, leading to errors, inconsistency, and subjectivity. According to IBISWorld, document preparation services in the US have a revenue of $5.6B and are expected to grow at a 1.9% CAGR between 2023 and 2028.

Intellectual Property

  • Early stage
Patent Information:
Category(s):
All
For Information, Contact:
Zachary Miles
Associate Vice President for Technology & Partnerships
University of Nevada, Las Vegas
702-895-4507
zach.miles@unlv.edu
Inventors:
Jee Woong Park
Cristian Arteaga-Sanchez
Keywords:
Computer Science & Software - Apps
Computer Science & Software - Software
Engineering & Manufacturing - Transportation
© 2024. All Rights Reserved. Powered by Inteum