Special Issue on VIII Brazilian Symposium on Neural Networks

In this Special Issue of the Journal of International Fuzzy Systems, JIFS, nine papers selected among the best evaluated accepted papers for the VIII Brazilian Symposium on Neural Networks, SBRN’04 are presented. This symposium covers topics related to Artificial Neural Networks, Evolutionary Computation, Fuzzy Systems and other Computational, accepting papers presenting new theoretical studies and novel applications. SBRN has an international Program Committee, with well known international researchers. In the last editions, the SBRN proceedings have been published by IEEE Computer Society. SBRN’2004 was held in the city of S̃ ao Lúıs, Brazil, between September 29th and October 1st. It was sponsored by the Brazilian Computer Society (SBC) and cosponsored by SIG/INNS/Brazil Special Interest Group of the International Neural Networks Society in Brazil. SBRN’04 received 329 submissions from several countries. Among these submissions, 154 full papers were accepted. From these papers, 22 high quality were preselected and their authors were asked to submit an extended and updated version for this special issue. The selection process took into account the originality, relevance and technical contribution. The new versions were submitted to a rigorous peer review process conducted by international reviewers. Nine papers recommended by the reviewers were accepted for this special issue. The accepted papers include both theoretical and application research works from very diverse areas, such as Neurophysiology, Neural Modeling, Hybrid Intelligent Systems, VLSI Implementation, Classification, Clustering, Associative Memory and Kernel Machines. Next, we briefly comment the topics covered by the papers published in this issue. The paperApplying genetic algorithms and SVMs to the gene selection problem, by Souza et al. describes a novel hybrid approach for gene selection from microarrays, based on Genetic Algorithms and Support Vector Machines (SVMs). The main idea of the paper is to use SVMs to estimate fitness in a gene selection strategy for the classification of micro-array tissue samples. Another paper also related to the combination of evolutionary strategies with SVMs is Evolutionary design of multiclass SVMs applied to protein structural class prediction, by Lorena et al. The idea is this paper is to use Genetic Algorithms to determine binary decompositions of multi-class problems. SVMs are then used in the classification of each one of the 2-class decomposed problems. Both papers are applied to Bioinformatics problems. Another hybrid intelligent system, combining neural networks and fuzzy systems is described in the paper Hybrid neural systems for large scale credit risk assessment applications, whose authors are Amorim and colleagues. In this paper, the authors compare the performance of two neuro-fuzzy models: Feature-weighted detector and fuzzy neural network in a large scale finance problem, credit risk analysis. They investigate the accuracy and the quality of the knowledge extracted by these models. The also use multi-layer perceptron networks in the comparison. Two papers on biological neurons were also selected for this special issue. The paper Bottom-up design of a Class 2 silicon nerve membrane, by Kohno and Aihara, is oriented to VLSI implementation. The authors extend their previous work in which they proposed a biologically realistic MOSFET-based Class 2 silicon nerve membrane to focus on a method of designing such a silicon structure. The approach is based on mathematical analyses that have been applied to biological neuron models. The other paper covering biological neurons isCalcium does not change memory in single calcium-activated potassium channel kinetics,