A task ontology model for domain independent dialogue management

Abstract

Dialogue systems have been a rapidly growing area in both scientific research and commercial application since 1990s. They can be applied in various fields, such as business, healthcare and education, etc. Due to its complexity, the design and development of a dialogue system is time consuming and costly. It is highly desirable for a generic dialogue system, especially dialogue management that is independent of specific domains. Methods or architecture for domain independent dialogue systems have been proposed by previous research in literature, however each of them has its own limitations and none has been widely adopted. This paper presents a new approach, a task ontology model for domain independent dialogue management. An abstract task ontology model is developed and based on this model a generic dialogue manager is created. Knowledge about a specific task is modeled in its task ontology and retrieved by an ontology reasoning component situated in the dialogue manager. Thus the dialogue manager is task or domain independent. A dialogue system is developed based on the proposed method and experimented with two different tasks: the book borrowing and the online train ticket booking. The experiment results indicate that the dialogue system can be readily applied to tasks from different domains without any modification. This paper has implications on future research and development of domain independent dialogue systems. It also contributes to the knowledge and dialogue system reuse and will have impact on the application of dialogue systems in a wider range of areas. v DEDICATION This thesis is dedicated to my dear husband and daughter for their endless love and support. vi ACKNOWLEDGEMENTS I would like to express my sincere gratitude to my supervisor, Dr. Xiaobu Yuan for his patience, enthusiasm and instruction on my study and research in the long journey towards the Master's Degree in Computer Science. Without his guidance, I could not image accomplishing this thesis. I also want to express my appreciation to Dr. Morrissey for her insightful comments and encouragement throughout the work. In addition, I would like to thank my classmates and friends who kindly helped and encouraged me during the completion of the thesis as well as my colleagues, especially Gwen and Cathy for their continuous support on my study. vii

A Task-Oriented Knowledge Base for Geospatial Problem-Solving

“…Based on the proposed architecture of the task-oriented knowledgebase described in Section 4.2, we build different abstract ontologies to represent the hierarchy and relationships of the concepts using Protégé 5.2.0 which is an OWL ontology development platform that allows creation and query of ontologies [21]. In general, an ontology is composed of the following components: concepts and properties of each concept, relationships or constraints between concepts, and instances of concepts [28]. Figure 8a presents all concepts or classes defined in the ontological knowledge base.…”

“…A case study of flood analysis is used to illustrate the effect and role of the task. Liu [28] proposed a task ontology model for domain-independent dialogue management and created a dialogue manager that is task-independent. Park et al [29] presented a task ontology based on travelers' perspectives using tasks, activities, relations, and properties.…”

Dialogue Management based on Multi-domain Corpus

“…(Annemiek, 2012) analyzes current dialogue management in operating unmanned systems and develops a more advanced way of dialogue management and accompanying dialogue manager. (Yuan, 2012) proposes a task ontology model for domain independent dialogue management, where the knowledge of a specific task is modeled in its task ontology which is independent from dialogue control. (Daubigney, 2012) proposes to apply the Kalman Temporal Differences (KTD) framework to the problem of dialogue strategy optimization so as to address all these issues in a comprehensive manner with a single framework.…”