Honors Program

 

Date of this Version

5-2019

Document Type

Thesis

Citation

Harrison, H., Onnen, J., Fitchett, L., Harkendorff, J., & Abolt, J. (2019.) Practical Natural Language Generation from Knowledge Graphs. Undergraduate Honors Thesis. University of Nebraska-Lincoln.

Comments

Copyright Henry Harrison, James Onnen, Lee Fitchett, John Harkendorff, and Joe Abolt 2019

Abstract

This project has two components, the Natural Language Generation (NLG) Engine and the Web App Editor. The NLG Engine can take a predefined knowledge graph (a method of representing connections between data points) as input, as well as accompanying data and variables, and using them to create meaningful and correct English sentences in order to find and convey the most important information to the end user. The Web App Editor can create and configure the knowledge graphs the Engine takes as input. It provides a clear web-based user interface that can create data points, variables, and scripts, as well as prototyping final values for the variables and editing the scripts. The Web App Editor also sends the template to the NLG Engine as it is being configured, so that the user can see the result of their template and observe the results of changes as they occur.

Share

COinS