人工智慧領域中非常有名的Turing Test給予本研究的第一個動機：一個專家系統不應該單單具備固定的解題知識；它也應該如同人類專家一樣，可以透過改採不同的觀點或解題策略來解題。本研究中我們試圖引進語意網表示的知識調整機制來增強專家系統的自我調適能力—專家系統因此能夠改用不同的觀點或解題策略來排除解題過程中遇到的困難。
傳統專家系統使用內建的解題知識及使用者提供的事實來求解，但這個事先建構的解題知識往往只能在事先假設的環境下運作，當使用者無法提供預期的輸入資料時，傳統專家系統即可能無法找到任何有用的解答。本研究中，我們在專家系統的推演過程中導入「自我調適(self-adaptability)」的觀念，並且提出一個能有效處理未預期情況的專家系統架構，讓專家系統可以在推理過程隨時組織新的操作知識(operational knowledge)來適應新的環境。藉由這種自我調適的特性，專家系統甚至在輸入資料不完全的情況下，仍然可以找到一個有用的解答。
我們所提出來的系統架構中，利用語意網路(semantic network)來表示領域知識(domain knowledge)。領域知識由三大部份所組成，分別為條件知識(condition knowledge)、結論知識(conclusion knowledge)、以及一組用來鏈接條件知識與結論知識的屬性關連(attribute relation)。我們定義了四種在領域知識上的調適(adaptation)操作，分別命名為條件知識調適(condition knowledge adaptation)、操作知識調適(operational knowledge adaptation)、結論知識調適(conclusion knowledge adaptation)、及表現方式調適(presentation adaptation)。本研究專注於探討前三種知識調適方式對專家系統調適能力的影響。另外，我們也提出了專家系統內建知識庫與外在語意詞庫合作的觀念，並藉由這種合作機制來擴展專家系統的調適能力，這樣也可藉由外在語意詞庫來減少知識工程師在建構專家系統知識庫時的工作量。
為了實現前述的知識調適能力，我們定義出兩個最基本的知識調適操作，分別為一般化知識調適(generalization knowledge adaptation)及特殊化知識調適(specialization knowledge adaptation)。此外，我們同時也根據系統在推演過程中是否需要使用者的指引，提出了兩種不同的調適策略，分別為監督式調適策略(supervised adaptation policy)及非監督式調適策略(unsupervised adaptation policy)。非監督式調適策略模式適合對問題背景知識較不熟悉的使用者，在此模式下，專家系統僅完全依靠內建知識庫所包含的各種觀念及相互間的層次關係來自動地進行知識調適的操作。另一方面，對一個具有問題背景知識的進階使用者來說，他可以利用監督式調適策略，在推理過程中提供適當的資訊來指引知識調適操作的進行，以求能快速地找到一個正確的解答。我們分別為這兩種調適策略定義了兩種不同的知識調適操作，分別命名為監督式知識調適及非監督式知識調適。
此外，為了讓自我調適式的專家系統能有效地找到一個好解答，我們提出了兩個以亂度(entropy)為基礎的量測：一個量測是用來在眾多可能的知識調適操作中，找出造成最小資訊損失(information loss)的操作；另一個量測則是讓系統在產生操作知識時，可以從眾多屬性關連(attribute relation)中找到最佳者。
我們還證明出，具有我們所提架構的自我調適式專家系統，不只在一般情況下可以找到一個正規的解答，在面臨未預期情況時，也可以藉由知識調適能力來自動地調適其操作知識並找到一個有意義的解答。最後，我們還透過一個精簡的例子來進一步說明自我調適式專家系統的運作方式。

The famous Turing Test gives the first inspiration to this research: an expert system should not be hard-wired with problem-solving knowledge; it should be able to exhibit the problem-solving capability like a human expert. In this study, we try to improve the self-adaptability of an expert system by regulating semantic-network-represented knowledge so that it can exhibit human-like behavior by taking flexible viewpoints to solve problems.
The pre-built knowledge of traditional expert systems is only capable of limited responses to changes in the operating environment. If the data input is unexpected, a traditional system may fail to reach any rational conclusions. In our study, we introduce the concept of self-adaptability to the inference process of an expert system, and propose an architecture that is capable of handling unexpected user input effectively and efficiently. Such a system can formulate operational knowledge on the move for inference. With this self-adaptive capability, an expert system can reach useful conclusions, even when the input data is insufficient.
The architecture of the proposed system encodes domain knowledge with semantic networks, which contain conclusion knowledge, condition knowledge, and attribute relations that relate the two types of knowledge together. We also define four types of adaptation, namely, condition knowledge adaptation, operational knowledge adaptation, conclusion knowledge adaptation, and presentation adaptation, and focus on how the first three contribute to the adaptive capability of the system. We also proposed the concept of the cooperation of the built-in domain knowledge and external semantic lexicon to extend the adaptability of the system and reduce the workload of a knowledge engineer in constructing domain knowledge.
We have defined two primary knowledge adaptation operations for realizing the adaptation mechanism in a self-adaptive expert system, which are generalization knowledge adaptation, and specialization knowledge adaptation. In addition, based on whether the assistance of the end-user advice is needed during the knowledge adaptation process, two adaptation policies are proposed, namely, supervised adaptation policy and unsupervised adaptation policy. The unsupervised adaptation policy mode is for an unskilled end-user, which allows the system to fully automatically perform knowledge adaptation based on the internal concept hierarchies of built-in domain knowledge. On the other hand, a proficient end-user can direct the process of knowledge adaptation under the supervised adaptation policy mode in order to find a useful conclusion quickly. We have also defined two corresponding types of knowledge adaptation operations for different adaptation policies, namely, unsupervised knowledge adaptation, and supervised knowledge adaptation.
In addition, to enable a self-adaptive expert system to effectively produce better conclusions, two entropy-based measuring mechanisms are proposed: one minimizes the information loss during knowledge adaptation, while the other selects the best attribute relation during the generation of operational knowledge.
We have proved that a self-adaptive expert system based on this architecture can always reach a regular conclusion or an abstract conclusion, which is a more meaningful conclusion by automatically modifying its operational knowledge in response to user feedback during the inference process, even in unexpected situations. Finally, we have demonstrated a small but clear example to illustrate how a self-adaptive expert system works.

[1] K. Aberer, P. Cudre-Mauroux, M. Hauswirth, and T. V. Pelt. “GridVine: Building Internet-Scale Semantic Overlay Networks,” Proceedings of International Semantic Web Conference (ISWC), pp. 107~121, Hiroshima, Japan, Nov. 7~11, 2004.
[2] P. J. Angeline. Adaptive and self-adaptive evolutionary computations. In M. Palaniswami and Y. Attikiouzel (Eds.), Computational Intelligence: A Dynamic Systems Perspective, pp. 152~163, IEEE Press, 1995
[3] S. Meyer-Nieberg, and H. G. Beyer, Self-Adaptation in Evolutionary Algorithms. In F. Lobo, C. Lima, and Z. Michalewicz (Eds.), Parameter Setting in Evolutionary Algorithm, pp. 47~75, Springer, Berlin, 2007.
[4] T.J.M. Bench-Capon, “The Role of Ontologies in the Verification and Validation of Knowledge Based Systems,” International Journal of Intelligent Systems, vol. 16. no. 3, pp. 377~391, 2001.
[5] A. Billig, E. Blomqvist, and F. Lin, “Semantic Matching Based on Enterprise Ontologies,” Proccedings of International Conference on Ontologies, DataBases, and Applications of Semantics (OTM2007), pp. 1161~1168, Nov. 25~30, 2007
[6] D. Billsus, C. A. Brunk, C. Evans, B. Gladish, and M. Pazzani, “Adaptive interfaces for ubiquitous web access,” Communications of the ACM, vol. 45, no. 5, pp. 34~38, May 2002.
[7] P. P. Bonissone, R. Subbu, K. S. Aggour, “Evolutionary optimization of fuzzy decision systems for automated insurance underwriting,” Proceedings of IEEE International Conference on Fuzzy Systems (FUZZ-IEEE’02), pp. 1003~1008, 2002.
[8] R. C. Chen, Y. C. Lee, and R. H. Pan, “Adding new concepts on the domain ontology based on semantic similarity,” Proceedings of International Conference on Business and Information, vol. 3, Singapore, July 12~14, 2006.
[9] L. Chittaro, and R. Ranon, “Hierarchical Model-Based Diagnosis based on Structural Abstraction,” Artificial Intelligence, vol. 155, no. 1-2, pp.147~182, 2004.
[10] S. H. Chung, J. V. Eepoel, and B. C. Williams, “Improving Model-based Mode Estimation through Offline Compilation,” International Symposium on Artificial Intelligence, Robotics and Automation in Space, St-Hubert, Canada, June 2001.
[11] P. Cobb, E. S. Yager, and C. Jacobus, “Anytime Diagnosis Using Model-Based Methods for Satellite Diagnostics,” Proceedings of the Twelfth International Florida Artificial Intelligence Research Society Conference, pp. 56~63, 1999.
[12] T. L. Dean, and M. Boddy, “An analysis of time-dependent planning,” Proceedings of the Seventh National Conference on Artificial Intelligence, pp. 49~54, 1988.
[13] S. M. Duggal, and R. Chhabra, “Learning Systems and Their Applications: Future of Strategic Expert System,” Issues in Information Systems, vol. 3, pp. 179~185, 2002.
[14] J. Euzenat, and P. Valtchev, “Similarity-based ontology alignment in OWL-lite,” Proceedings of the European Conference on Artificial Intelligence (ECAI), pp. 333~337, 2004.
[15] T. R. Gruber, “Toward Principles for the Design of Ontologies used for Knowledge Sharing,” International Journal of Human and Computer Studies, vol. 43, no. 5/6, pp. 907~928, 1995.
[16] J. W. Grzymala-Busse, “Rough set strategies to data with missing attribute values,” Proceedings of the Workshop on Foundations and New Directions in Data Mining, associated with the third IEEE International Conference on Data Mining, pp. 56~63, November 19~22, 2003.
[17] J. W. Grzymala-Busse, and M. Hu “A Comparison of Several Approaches to Missing Attribute Values in Data Mining,” Proceedings of the Second International Conference on Rough Sets and Current Trends in Computing, RSCTC 2000, vol. 2005 of Lecture Notes in Computer Science, Springer, 2001.
[18] J. N. D. Gupta, S. K. Sharma, and J. Hsu, “An overview of knowledge management,” In J. N. D. Gupta, and S. K. Sharma (Eds.), Creating knowledge based organizations, pp. 1~28, UK: Idea Group Publishing, 2004.
[19] Y. Kalfoglou, and M. Schorlemmer, “Ontology mapping: the state of the art,” The Knowledge Engineering Review, vol. 18, no.1, pp. 1~31, Jan., 2003.
[20] G. Karsai, and J. Sztipanovits, “A model-based approach to self-adaptive software,” IEEE Intelligent Systems, vol. 14, no. 3, pp. 46~53, May/Jun, 1999.
[21] J. S. Kim, and S. Park, “Self Adaptive Software Technology for Robotics.” Proceedings of the 11th Asia-Pacific Software Engineering Conference, pp.740~741, Nov. 30~Dec. 03, 2004.
[22] Y. Kitamura, and R. Mizoguchi, “Ontology-based systematization of functional knowledge,” Journal of Engineering Design, Taylor & Francis, vol. 15, no. 4, pp. 327~351, August 2004.
[23] R. Laddaga, and P. Robertson, “Model Based Diagnosis in Self Adaptive Software,” Proceedings of International Workshop on Principles of Diagnosis, Washington, DC, USA, June 11~14, 2003.
[24] L. Laera, V. Tamma, T. Bench-Capon and, G. Semeraro, “SweetProlog: A System to Integrate Ontologies and Rules,” In G. Antoniou and H. Boley, Rules and Rule Markup Languages for the Semantic Web: Third International Workshop, RuleML 2004, Lecture Notes in Computer Science 3323, pp. 188~193, Springer:Berlin, 2004.
[25] R. Laddaga, and P. Robertson, “Self Adaptive Software: A Position Paper,” Proceedings of International Workshop on Self-* Properties in Complex Information Systems, Bertinoro, Italy, 2004.
[26] A. Lekovaa, and D. Batanov, “Self-testing and self-learning fuzzy expert system for technological process control,” Computers in Industry, vol. 37, no. 2, pp. 135~141, 1998.
[27] L. Lhotska, and T. Vlcek, “Efficiency enhancement of rule-based expert systems,” Proceedings of the 15th IEEE Symposium on Computer-Based Medical Systems (CBMS 2002), pp. 53~58, June 4~7, 2002.
[28] M. Li, X. Wang, and Z. Wang, “Two Language Models Using Chinese Semantic Parsing,” Tsinghua Science & Technology, vol. 11, no. 5, pp. 582~588, Oct. 2006.
[29] Y. T. Lin, S. S. Tseng, and C. F. Tsai “Design and implementation of new object-oriented rule base management system,” Expert Systems with Applications, vol. 25, no. 3, pp. 369~385, 2003.
[30] H. Liu, H. Bao, and J. Feng, “Instance assisted ontology alignment for digital museums,” Proceedings of the 7th WSEAS international Conference on Simulation, Modelling and Optimization, pp. 79~84, Beijing, China, Sept. 15~27, 2007.
[31] J. Liu, C. K. Wong, and K. K. Hui, “An Adaptive User Interface Based On Personalized Learning,” IEEE Intelligent Systems, vol. 18, no. 2, pp. 52~57, May 2003.
[32] L. J. Liu, Y. D. Wang, and M. Z. Guo, “The research and application of the self-learning expert system based on BP network,” Proceedings of International Conference on Machine Learning and Cybernetics, vol. 7, pp. 4153~4157, Aug. 2005.
[33] G. A. Miller, “WordNet: A Lexical Database,” Communications of the ACM, vol. 38, no. 11, pp. 39~41, Nov. 1995.
[34] G. A. Miller, R. Beckwith, C. Fellbaum, D. Gross, and K. J. Miller, “Introduction to WordNet: An On-line Lexical Database,” International Journal of Lexicography, vol. 3, no. 4, pp. 235~244, 1990.
[35] D. Milward, and M. Beveridge, “Ontology-based dialogue systems,” Proceedings of IJCAI WS on Knowledge and reasoning in practical dialogue systems, Acapulco, Mexico, August 10, 2003.
[36] N. F. Noy, and D. McGuinness, “Ontology Development 101: A Guide to Creating Your First Ontologym,” Stanford Knowledge Systems Laboratory Technical Report KSL-01-05 and Stanford Medical Informatics Technical Report SMI-2001-0880, March 2001.
[37] E. Oreizy, M. M. Gorlick, R. N. Taylor, D. Heimbigner, G. Johnson, N. Medvidovic, A. Quilici, D. S. Rosenblum, and A. L. Wolf. “An Architecture-Based Approach to Self-Adaptive Software,” IEEE Intelligent Systems, vol. 14, no. 3, pp. 54~62, May/June 1999.
[38] X. Pan, “An Experimental Adaptive Expert System,” Proceedings of InterSymp-2000, Baden-Baden, Germany, July 31~August 4, 2000.
[39] M. Pantic, and L. J. M. Rothkrantz, “Self-adaptive expert system for facial expression analysis,” Proceedings of IEEE SMC, pp. 73~79, 2000.
[40] J. Prentzas, I. Hatzilygeroudis, and J. D. Garofalakis, “A Web-Based Intelligent Tutoring System Using Hybrid Rules as Its Representational Basis,” Proceedings of the 6th international Conference on intelligent Tutoring Systems, vol. 2363, Springer-Verlag, London, pp. 119~128. June 2~7, 2002.
[41] Z. Qin, C. X. Ling, and S. Sheng, “Missing Is Useful: Missing Values in Cost-Sensitive Decision Trees,” IEEE Transactions on Knowledge and Data Engineering, vol. 17, no. 12, pp. 1689~1693, 2005.
[42] J. R. Quinlan, “Induction of Decision Trees,” Machine Learning,vol. 1, no. 1, pp. 81~106, 1986.
[43] J.R.Quinlan, C4.5: Programs for Machine Learning, Morgan Kaufmann, San Mateo CA, 1993.
[44] D. Richards, “Ripple down rules: a technique for acquiring knowledge,” In Decision Making Support Systems: Achievements, Trends and Challenges For, M. Mora, G. A. Forgionne, and J. N. Gupta, Eds. Idea Group Publishing, Hershey, PA, pp. 207~226, 2003.
[45] P. Robertson, and R. Laddaga, “Model Based Diagnosis and Contexts in Self Adaptive Software,” In Ozalp Babaoglu, Márk Jelasity, Alberto Montresor, Christof Fetzer, Stefano Leonardi, Aad van Moorsel, and Maarten van Steen (Eds.), Self-Star Properties in Complex Information Systems, volume 3460 of Lecture Notes in Computer Science, Hot Topics, pp. 112~127, Springer-Verlag, 2005.
[46] D. W. Rolston, Principles of Artificial Intelligence and Expert Systems Development, McGraw-Hill, Singapore, 1988.
[47] E. S. Sazonov, P. Klinkhachorn, H. V. S. Gangarao, U. B. Halabe, “Fuzzy logic expert Ssystem for automated damage detection from changes in strain energy mode shapes,” Nondestructive Testing and Evaluation. vol. 18, no. 1, pp. 1~17, 2002
[48] C. E. Shannon, “A mathematical theory of communication,” Bell Syst. Tech. Journal., vol. 27, pp. 379~423, 1948.
[49] B. Smyth, K. Bradley, and R. Rafter, “Personalization techniques for online recruitment services,” Communications of the ACM, vol. 45, no. 5, pp. 39~40, May 2002.
[50] P. Spirtes, “An anytime algorithm for causal inference,” Proceedings of AISTATS, pp. 213~231, San Francisco: Morgan Kaufmann, 2001.
[51] V. Sugumaran, and V. C. Storey, “Ontologies for conceptual modeling: their creation, use, and management,” Data & Knowledge Engineering, vol. 42, no. 3, pp. 251~271, Sept. 2002.
[52] S. Thiel, S. DalPKAkis, and D. Roller, “A Learning Agent for Knowledge Extraction from an Active Semantic Network,” Proceedings of World Academy of Science, Engineering and Technology, vol. 7, pp. 217~220, Aug. 2005.
[53] C. Tran, A. Abraham, and L. Jain, “TACDSS: Adaptation Using a Hybrid Neuro-Fuzzy System,” Proceedings of the Seventh Online World Conference on Soft Computing in Industrial Applications (WSC7), Advances in Soft Computing: Engineering Design and Manufacturing, Benitez, J. M. et al (Eds.), Springer Verlag, Germany, pp. 53~62, 2002.
[54] E. Turban, Expert System and Applied Artificial Intelligence. New York: Macmillan Publishing Company. 1992.
[55] A. M. Turing, “Computing Machinery and Intelligence,” Mind, vol. 59, no. 236, pp. 433~460, 1950.
[56] A. Verberne, F. V. Harmelen, and A. T. Teije, “Anytime Diagnostic Reasoning using Approximate Boolean Constraint Propagation,” Proceedings of the Seventh International Conference on Principles of Knowledge Representation and Reasoning, pp. 323~332, 2000.
[57] L. Vizenor, O. Bodenreider, L. Peters, A. T. Mccray, “Enhancing biomedical ontologies through alignment of semantic relationships: Exploratory approaches,” Proceedings of AMIA Annual Symposium, pp. 804~808, 2006
[58] Q. Wang, “Towards a rule model for self-adaptive software,” SIGSOFT Software Engineering Notes, vol. 30, no. 1, Jan. 2005.
[59] W. Wang, and R. Rada, “Structured Hypertext with Domain Semantics,” ACM Transactions on Information Systems, vol. 16, no. 4, pp. 372~412, Oct. 1998.
[60] X. W. Wang, H. B. Qu, P. Liu, and Y. Y. Cheng, “A self-learning expert system for diagnosis in traditional Chinese medicine,” Expert Systems with Applications, vol.26, pp. 557~566, 2004.
[61] C. Wanger, “End Users as Expert System Developers?” Journal of End User Computing, vol. 12, no. 3, pp. 3~13, 2000
[62] A. Wichert, “Categorial Expert Systems,” Expert Systems, vol. 21, no. 1, pp. 34~47, 2004.
[63] S. Wiriyacoonkasem, and A. C. Esterline, “Adaptive learning expert systems,” Proceedings of the IEEE Southeastcon 2000, pp. 445~448, 2000.
[64] M .E. Yahia, R. Mahmod, N. Sulaiman, and F. Ahmad, “Rough neural expert systems,” Expert Systems with Applications, vol. 18, pp. 87~99, 2000.
[65] J. Zhang, A. Silvescu, and V. Honavar, “Ontology-Driven Induction of Decision Trees at Multiple Levels of Abstraction,” Proceedings of Symposium on Abstraction, Reformulation, and Approximation, pp. 316~323, Springer-Verlag, 2002.
[66] S. Zilberstein, “Using Anytime Algorithms in Intelligent Systems,” AI Magazine, vol. 17, no. 3, pp. 73~83, 1996.
[67] S. Zilberstein, and S. Russell, “Approximate Reasoning Using Anytime Algorithms,” In S. Natarajan (Ed.), Imprecise and Approximate Computation, Kluwer Academic Publishers, 1995.