| Understanding the decision-making processes of living systems, and the efforts to mimic them, are identified with research in Artificial Intelligence (AI). However, the recent popularity of neural networks, fuzzy systems and evolutionary computation, which are widely considered as areas related in AI, has created a need for a new definition to distinguish them from traditional AI techniques.
Lotfi Zadeh, the inventor of fuzzy logic, has suggested the term “Soft Computing.” He created the Berkeley Initiative of Soft Computing (BISC) to connect researchers working in these new areas of AI. In contrast to hard computing, soft computing techniques account for the possibility of imprecision, uncertainty and partial truth.
The first joint conference of neural networks, fuzzy systems, and evolutionary computation organized by the Institute of Electrical and Electronic Engineers (IEEE) in 1994 was the World Congress on Computational Intelligence. In his paper at this historic joint conference, James Bezdek defined three kinds of intelligence to distinguish Computational Intelligence from traditional AI and living systems: biological or organic, artificial or symbolic, and computational or numeric. All natural creatures belong to the first category, while traditional AI techniques remain in the second category. Computational Intelligence (CI) is the group of techniques based on numerical or sub-symbolic computation mimicking the products of nature. However a number of “new AI” methods have found a home in CI. The application of CI in emerging research areas, such as Granular Computing, Mechatronics, and Bioinformatics, show its maturity and usefulness. Recently IEEE changed the name of its Neural Network society to become the IEEE Computational Intelligence Society.
The International Conference on Fuzzy Systems and Knowledge Discovery (FSKD) held in Singapore in 2002, in conjunction with two other conferences in CI, has led to the publication of these two edited volumes. This volume contains CI methods and applications in modeling, optimisation and prediction. The other volume entitled “Classification and Clustering for Knowledge Discovery”, from the same publisher, includes the papers on Clustering and Classification. |