The Society Library
NLP Engine for Argument Detection
An innovative NLP engine that enhances article analysis by identifying logical arguments.
About the Project
The Argument Miner project was initiated in collaboration with The Society Library for The Great American Debate project, to enhance the analysis of written content. By leveraging the ArgumenText API, we aimed to create a system that could effectively parse articles and identify logical arguments made by authors. This project was driven by the need for better tools in understanding complex texts and the claims they present.
Our approach involved evalutating the existing the NLP model to ensure it accurately detected various types of arguments, which included premises, conclusions, and supporting evidence, while maintaining reference metadata.
This project served as a launching point for more in depth tools that helped create The Society Library's platform for tracking public discourse on important topics.
Project Execution
An evaluation framework to test the current state of the art in argument detection was a required launching point for The Great American Debate to be good at its intentions. Leveraging existing architecture for a large corpus of articles, I was able to extract and evaluate the efficacy of the argument detection model for tracking debate arguments.
• Developed the core NLP engine using the ArgumenText API for argument detection
• Conducted extensive testing to evaluate and refine the model's accuracy and performance
• Collaborated with The Society Library to integrate the engine into their platform
