DMCA
Three Lessons in Creating a Knowledge Base to Enable Reasoning, Explanation and Dialog
| Citations: | 1 - 1 self |
BibTeX
@MISC{Chaudhri_threelessons,
author = {Vinay K. Chaudhri and Nikhil Dinesh and Daniela Inclezan},
title = {Three Lessons in Creating a Knowledge Base to Enable Reasoning, Explanation and Dialog},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract. Our work is driven by the hypothesis that for a program to answer questions, explain the answers, and engage in a dialog just like a human does, it must have an explicit representation of knowledge. Such explicit representations occur naturally in many situations such as engineering designs created by engineers, a software requirement created in unified modeling language or a process flow diagram for a manufacturing process. Automated approaches based on natural language processing have progressed on tasks such as named entity recognition, fact extraction and relation learning. Use of automated methods can be problematic in situations where the conceptual distinctions used by humans for reasoning are not directly expressed in natural language or when the representation must be used to drive a high fidelity simulation. In this paper, we report on our effort to systematically curate a knowledge base for substantial fraction of text in a biology textbook [26]. While this experience and the process is interesting on its own, three aspects can be especially instructive for future development of knowledge bases by both manual and automatic methods: (1) Consider imposing a simplifying abstract structure on natural language sentences so that the surface form is closer to the target logical form to be extracted. (2) Adopt an upper ontology that is strongly motivated and influenced by natural language. (3) Develop a set of guidelines that captures how the conceptual distinctions in the ontology may be realized in natural language. Since the representation created by this process has been quite effective for answering questions and producing explanations, it gives a concrete target for what information should be extracted by the automated methods.
Keyphrases
knowledge base natural language explicit representation conceptual distinction automated method natural language sentence entity recognition manufacturing process concrete target simplifying abstract structure relation learning substantial fraction automatic method software requirement unified modeling language future development natural language processing many situation surface form fact extraction biology textbook process flow diagram high fidelity simulation automated approach target logical form upper ontology engineering design
