Abstract
Structured information is information that computers can easily understand and reason with, such as algebraic expressions, logical formulas, frames, and database tables. Unstructured information is information that is not represented in such a form and is difficult for computers to understand. Examples of unstructured information include natural language text, audio, images, and video. Because unstructured information is so easy for humans to create, it is abundant. Much of the world’s data is unstructured. Given that there is so much unstructured information available, the question arises of how we might be able to use it for automated commonsense reasoning. We discuss the use of unstructured information, specifically natural language text, for commonsense reasoning. First, we describe attempts to use natural language as a programming language. We then review several restricted versions of English for reasoning, including reasoning in the event calculus. Then we discuss automated reasoning directly using natural language. Next, we describe the Watson system for natural language question answering and the WatsonPaths system built on top of Watson, which use unstructured information as a source of knowledge for reasoning. Finally, we compare the various natural language reasoning systems and methods.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call