Since introduction of the firsts Artificial Neural Networks (ANN) models in fifties and the designing of the earliest ANN based approaches in seventies, what is called, today, “Bio-Inspired Artificial Intelligence” (BIAI) has undergone a number of decisive steps becoming the new and entire science attracting a large number of nowadays scientific communities. If over the past decades, ANN and issued techniques have allowed the elaboration of many original techniques covering a large field of applications overcoming limitations of conventional approaches thank to their learning and generalization capabilities, they also made appear a number of expectations for designing “intelligent” information processing systems. However, if learning and generalization capabilities of ANN are among central features to take control of “intelligent information processing”, it is now admitted that “artificial intelligent behavior” will require more sophisticated mechanisms than those performed by these
“simple” models: in fact, the complexity of biological mechanisms doesn’t endorse to imagine emergence of some “artificial intelligent behavior” comparable to the performances accomplished by alive systems. Even if much is still unknown about how the brain trains and selforganizes itself to process so complex and so diverse information, the recent advances in biology and especially in “neurobiology” allowed highlighting some of key mechanisms of biological intelligence. Among them, brain’s structure’s “modularity”, its “self-organizing” capabilities and its complexity reduction ability seem federate convergent agreements. If it is still early to state on “concurrent” or “cooperative” nature of ways that these complex features interact, they are still considered as foremost supplies for higher level artificial intelligent behavior emergence. However, looking forward benefiting from further theoretical advances, new experimental discoveries and novel technological improvements elucidating brain’s structure and its operation enigmas, we can still take advantage from its most popular features (learning and generalization capabilities) as well from its other nameless but very appealing properties (natural parallelism, operation redundancy, distributive information storage) to improve actual computational capabilities and to open the road for intelligent processing of information. As for ANNIIP 2005 workshop, where the aim was to make an inventory of recent advances in ANN and Intelligent Information Processing areas around a deliberately limited number of presentations, the main goal of this book (issued from ANNIIP 2005 workshop) is to convene around an intentionally small number of topics a set of relevant papers concerning the aforementioned fields. The choice of a restricted number of topics has been motivated by the premeditated desire to surround more accurately a number of foremost purposes with regard to “ANN and Intelligent Information Processing” preventing a dispersive compilation of a larger number of matters. Four leading themes have been identified making the primary frame of the present collective volume. The first one concerns learning, architecture and optimization of ANN. The second one called “hybrid and multi agent approaches” surrounds aspects related to hybridization, cooperation and modularization. A particular interest has been devoted to applicative aspects with a special attention on “real-world and industrial” applications. The last selected theme deals with “fuzzy and neuro-fuzzy” approaches, another key research area related to artificial intelligent behavior. It is important to remind that scientific relevance and technical quality of a collective volume emerge from quality of its contributors: those who contribute by the high quality of their manuscripts and those who take part in reviewing of submitted papers ensuring the excellence of the book by their valuable expertise. That is why, I would like to acknowledge contributors of all accepted papers. For this same reason, I would like to express my gratitude to Reviewing Board and Program Committee for the valuable work that they accomplished. Finally, I would like to express my particular gratitude and my warm appreciation to my friend Prof. Joaquim Filipe, ICINCO 2005 Conference Chair, to convince and to trust me for chairing the ANNIIP 2005 workshop under the frame and with support of his valuable conference. Before concluding the present foreword, it is also essential to remind that organizing a conference as ICINCO, with three related workshops as ANNIIP 2005, is a challenging undertaking requiring reliable and solid Organizing Committee. I would like acknowledge all members of ICINCO 2005 Organizing Committee, with a special attention for Marina Carvalho from ICINCO Secretariat who was particularly involved in ANNIIP 2005 organization tasks.
Vol. 1 - 972-8865-36-8