| 2003 Publications |
|
| Weighted MCRDR: Deriving Information about Relationships between Classifications in MCRDR. |
|
| R. Dazeley and B. H. Kang
|
| The 16th Australian Joint Conference on Artificial Intelligence (AI?3). 2003. Perth, Australia: Springer-Verlag, Berlin Heidelberg New York.
|
| Abstract. Multiple Classification Ripple Down Rules (MCRDR) is a knowledge acquisition technique that produces representations, or knowledge maps, of a human expert knowledge of a particular domain. However, work on gaining an understanding of the knowledge acquired at a deeper meta-level or using the knowledge to derive new information is still in its infancy. This paper will introduce a technique called Weighted MCRDR (WM), which looks at deriving and learning information about the relationships between multiple classifications within MCRDR by calculating a meaningful rating for the task at hand. This is not intended to reduce the knowledge acquisition effort for the expert. Rather, it is attempting to use the knowledge received in the MCRDR knowledge map to derive additional information that can allow improvements in functionality of MCRDR in many problem domains. Preliminary testing shows that there exists a strong potential for WM to quickly and effectively learn meaningful weightings.
|
|
|
|
|
|
| Rated MCRDR: Finding non-Linear Relationships Between Classifications in MCRDR. |
|
| R. Dazeley and B. H. Kang
|
| 3rd International Conference on Hybrid Intelligent Systems. 2003. Melbourne, Australia: IOS Press.
|
| Abstract. Multiple Classification Ripple Down Rules (MCRDR) is a simple and effective knowledge acquisition technique that produces representations, or knowledge maps, of a human expert knowledge of a particular domain. This knowledge map can then be used to automate and help the user perform classification and categorisation of cases while still being able to add more refined knowledge incrementally. While MCRDR has been applied in many domains, work on understanding the meta-knowledge acquired or using the knowledge to derive new information is still in its infancy. This paper will introduce a technique called Rated MCRDR (RM), which looks at deriving and learning information about both linear and non-linear relationships between the multiple classifications within MCRDR. This method uses the knowledge received in the MCRDR knowledge map to derive additional information that allows improvements in functionality within existing domains, to which MCRDR is currently applied, as well as opening up the possibility of new problem domains. Preliminary testing shows that there exists a strong potential for RM to quickly and effectively learn meaningful ratings.
|
|
|
|
|
|
| Web Information Management System: Personalization and Generalization
|
|
| Sung Sik Park, Yang Sok Kim and Byeong Ho Kang
|
| IADIS International Conference WWW/Internet 2003, November 5-8, 2003,
Algarve Portugal
|
| Abstract. Our research focuses on web information management for people who want to monitor and use the World Wide Web (WWW) information, as their information resource. Web information is mainly open to the public and search engines are widely used, although there are complaints about the large amount of irrelevant information. Some web technology research focuses on promptness of changed information and relevance to users. Most web search engines do not fully satisfy these two requirements in general because they try to cover all web sites and users together. The aim of Web monitoring research is to overcome these problems. Web monitoring systems check predefined web pages and prompt users when there are changes in these pages. This research focuses on the management of newly uploaded information in target web sites. In a web monitoring system, the system should cover various different types of web pages (generalization), as well as personal aspects (personalization). This system achieves these tasks by integrating the first and third modules. In addition to web monitoring functions, the second module provides the information management functionality in the user¡¯s local storage.
|
|
|
|
|
|
| Using Multiple Classification Ripple Down Rules for Intelligent Tutoring System¡¯s Knowledge Acquisition
|
|
| Yang Sok Kim, Sung Sik Park, Byeong Ho Kang and Joa Sang Lim
|
| The 16th Australian Joint Conference On Aritificial Intelligient- AI'03, 2003, Perth, Western Australia |
| Abstract. This research focuses on the knowledge acquisition (KA) for an intelligent tutoring system (ITS). ITSs have been developed to provide considerable flexibility in presentation of learning materials and greater abilities to respond to individual students¡¯ needs. Our system aims to support experts who want to accumulate the classification knowledge. Rule based reasoning has been widely used in ITSs. Knowledge acquisition bottleneck is a major problem in ITSs as it is known in AI area. This problem hinders the diffusion of ITSs. MCRDR is a well known knowledge acquisition methodology and mainly used in classification domain. MCRDR is used to acquire knowledge for the classification of learning materials (objects). The new ITS is used to develop a part of online education system for the people who learn English as a second language. Our experiment results show that the classification of learning materials can be more flexible and can be organized in multiple contexts.
|
|
|
|
|
|
| Framework for Knowledge Acquisition within
Conversational Agents
|
|
| Pauline Mak, Byeong Ho-Kang and Mike Cameron-Jones
|
| Smart Internet Technology CRC
Annual Partner Conference, Sydney, 18 & 19 September, 2003 |
| Abstract. Most existing conversational agents are developed for specific domains and it is difficult to be applied to other domains or to be maintained. Such difficulties are primarily caused by a poor separation between domain specific knowledge in the agent and the general conversation. The preliminary framework described here separates the domain specific knowledge from the general conversation. To acquire and maintain the domain/personal knowledge, we have developed a knowledge acquisition module for the agent. Since RDR has been used successfully with several expert systems, this paper demonstrates an agent that is built over an existing conversational agent Probot that utililses Ripple Down Rules (RDR), a knowledge acquisition method.
|
|
|
