It can be observed from the previous section that Computational Intelligence (CI) methods cover a very broad area of research, and it is not easy to summarize what are these areas. Based on the purpose of CI, on the scopes of CI, ML, TNN, NN, top tier Conferences, and on the past and present contributions of the author to the field, two main areas of computational learning systems can be identified:
This seminar presents a very brief state-of-the-art in these two areas. Several applications are presented for each area that result from the author’s work, together with many other researchers. Note however that a much larger range of applications in the field of computational learning systems can be found in the literature. This seminar is organized into 5 chapters. The outline of the chapters and the main topics that each chapter considers are presented in the following.
In Chapter 2 neural networks are presented, tackling three architectures (Multi-layer Perceptron, Modular Networks and Hopfield Networks) with distinct goals meanwhile pursuing the final purpose of providing solutions for complex problems.
Kernel methods, namely, Support Vector Machines (SVMs) are described in Chapter 3 spanning three mathematical formulations (Hard margin, Soft margin and Nonlinear approach).
In Chapter 4 applications derived from the models presented previously are briefly described.
This seminar ends with Chapter 5 where the future developments of the computational learning systems are discussed. Future work towards the developement of whitening models are pointed as further lines of research.