Question-answering dialogue system for emergency operations

By Hao-Yung Chan, Meng-Han Tsai. 2019

26 citations recorded by the source corpus.

Abstract This research aims to develop a dialogue system for eliminating difficulties encountered in the efficient and accurate utilization of information for decision makers at emergency operation centers (EOC). The rapid growth in the amount of data has caused complications in decision making tasks in disaster management. This has increased the difficulty for decision makers when they try to accomplish their mission accurately and efficiently. In our preliminary work, we attempted to develop a chatbot for decision makers and staff at an EOC to assist them in the utilization, selection, and processing of information efficiently and accurately. A user experience test revealed that a better adaption of the dialogue system to the characteristics of disaster management and the EOC was required. Thus, we specifically focused on the development of a technique of language understanding (LU) to be able to analyze the user's demands. As a solution, we developed a framework for a dialogue system combining question answering features, a knowledge base as the knowledge provider, and a search module that can handle the relatively difficult querying tasks. To improve the performance of the system, we focused on enhancing the capability in understanding users' questions. The construction of the question analysis and knowledge base were both adapted to the characteristics of disaster management. The validation shows that our method can analyze and process the user's questions into a machine-acceptable form with a success rate of approximately 70%.

Subfields: Artificial Intelligence, Question Answering & Reading Comprehension, Dialogue & Conversational AI