|Title:||Knowware System for the Development of KBS|
December 21, 2009 (Monday)
2:30 p.m. - 3:30 p.m.
Room 1027, 10/F, Ho Sin-hang Engineering Building,
The Chinese University of Hong Kong,
Prof. Liya Ding
Dean, Faculty of Information Technology
Macau University of Science and Technology
With the great achievements in research and development of intelligent techniques, it has attracted more attention that how to better appreciate those techniques and further promote the application of intelligent systems in wide areas. The success of development and application of an intelligent system requires the availability of two groups of people: AI experts who hold the techniques and tools for problem solving, and domain experts who know well the problem to be solved and hold domain knowledge leading to a necessity of the development of intelligent system. However, in reality, it is often a challenge to get the both groups working together to derive the inherent synergies.
A knowledge-based system (KBS) especially a hybrid intelligent system is, in some sense, a set of knowledge-based processing units inter-connected in an appropriate way to ensure a desired inference flow. Knowware System (KWS) has been proposed to support KBS developer by offering a set of intelligent components stored in KWS warehouse together with the function of automatic construction. KWS not only allows developers to easily design their system, but also realizes an automatic construction of the target KBS based on the developers' design. The user (KBS developer) can easily input his/her design ideas as well as domain knowledge, and deploy KWS to construct a customized KBS based on the information and knowledge sources provided. This concept is general, and not necessarily limited to any specific application.
The inference structure of a KBS constructed by KWS is represented as a knowledge hierarchy connecting multiple intelligent components associated with corresponding knowledge for problem solving. The task of construction of knowledge hierarchy can be done either through the interactive graphic editor or the processor of knowledge description language (KDL). Such a knowledge hierarchy represents statically the inference structure of the target KBS. In order to make it a complete stand-alone application, an inference engine monitors and controls the dynamic execution of KBS, including protocol management, confidence handling, and order of execution. Individual intelligent processors may process data, information and knowledge of different types, adopt different inference mechanisms that are suitable to inputs and outputs of varying nature.
This presentation will give a brief introduction to KWS: the KWS framework, KWS engine, intelligent components, knowledge description language (KDL) and its processor. The discussion will also cover the truth value flow inference in KBS constructed by KWS, confidence handling in individual components, and forward inference with possible partial feedback.
This work was supported partly by Macao Science and Technology Development Fund under grant 048/2006/A (2007.7~2009.7).
Liya Ding received her Bachelor of Engineering in Computer Technology from Shanghai University of Technology (now Shanghai University), China in Jan 1982; Ph.D. in Computer Science from Meiji University, Japan, Mar 1991. From 1991 to 2002, she was with the Institute of Systems Science (ISS), National University of Singapore (NUS), Singapore, being the project leader for Real World Computing (RWC) Partnership Neuro-ISS Lab 1993-1996, Senior Lecturer/Consultant, Programme Director for Knowledge Engineering, Master of Technology 1998-2002; From fall 2002 to 2005, was the Dean of the School of Intelligent Systems and Technology, Inter-University Institute of Macau (IIUM), Macau; From fall 2005 to present, is with Macau University of Science and Technology, being the Vice Dean / Dean of the Faculty of Information Technology.
Main research interests include: fuzzy inference and approximate reasoning, knowledge-based systems (KBS), knowledge engineering and soft computing. Was the editor of book "A New Paradigm of Knowledge Engineering by Soft Computing", World Scientific, 2001.
Enquiries: Miss Temmy So at tel 2609 8444
For more information, please refer to http://www.cse.cuhk.edu.hk/seminar