Research Project Title:
Linguistic Analysis of Wikipedia for Question Answering
abstract:Question-answering systems are becoming increasingly prevalent in modern society, largely in services such as Google Home or Amazon Echo. The two major components of a question-answering system are the question-parsing software and the knowledge base used to find answers. My research focuses on exploring ways to generate useful knowledge bases out of semi-structured data. In particular, the content of Wikipedia can be parsed and organized in order to efficiently answer a broad range of questions. The goal of my project is to add to WikipediaBase, a knowledge base used for the question-answering system START, by exploring different methods of information extraction and pattern recognition in the website. These methods will demonstrate new ways to add breadth and depth to generated knowledge bases.
“I am participating in SuperUROP to gain a deeper understanding of the academic process. While doing so, I also hope to make meaning contributions to fields of computer science that intrigue me. I have participated in UROPs every year since starting at MIT, which has prepared me to delve into a more ambitious research project. I am particularly excited about learning how to write professional academic papers.”