Neural Interface for Web-Scale Knowledge

Abstract

Modern natural language tasks are increasingly dependent on external world knowledge. My PhD study has particularly focused on three challenges in this literature: making sense of unstructured knowledge, leveraging extremely large knowledge, and reasoning over the knowledge data. I will mainly discuss my approaches to tackle these challenges and how they can serve as an effective interface for interacting with the world knowledge. I will conclude with an argument that designing a seamless and universal knowledge interface is a crucial research goal that can better address knowledge-dependency problem in machine learning tasks.