Zadehian Paradigms for Knowledge Extraction in Intelligent Manufacturing

Manufacturing is a knowledge-intensive activity. The knowledge underlying a specific manufacturing process or system is often extracted from a small set of experimental observations. To automate the knowledge extraction process various machine learning methods have been used (Pham & Afifi, 2005; Monostori, 2003). Even though such methods are used, a great deal of human intelligence (knowledge extractor’s judgment, preference) is required for getting good results (Ullah & Khalifa, 2006). As a result, a machine learning method that is able to utilize human cognition as straightforwardly as possible seems more realistic for extracting knowledge in manufacturing. In fact, human-assisted machine learning methods are in agreement with the modern concept of manufacturing automation—how to support humans with computers rather than how to replace humans by computers (Kals et al., 2004). Thus, for advanced manufacturing systems, the machine learning methods wherein humans and computers compliment each other and the course of knowleldge extraction is determine by the human cognition rather than by a fully automated algorethemic approach is desirable. Artificial intelligence community has also started to recognize the need for human-assisted machining learning methods (i.e., human comprehensible machine learning methods):