**Manuel Ojeda-Aciego** holds a MSc in Mathematics (1990) and a PhD in Computer Science (1996). He is currently Full Professor in the Department of Applied Mathematics, University of Málaga. He is the president of the Computer Science Committee of the Royal Spanish MathematicalSociety (2005), Area Editor of Fuzzy Fundamentals of the Intl J on Uncertainty and Fuzziness in Knowledge-based Systems, member of the Editorial Board of the IEEE Tr on Fuzzy Systems, member of the Steering Committee of the Intl Conf onConcept Lattices and their Applications (CLA) and the Intl Conf on Information Processing and Management of Uncertainty in knowledge-based systems (IPMU), member of EUSFLAT, and senior member of the IEEE. As a member of the Research Group of Applied Mathematics in Computing, his current research interests include residuated and multi-adjoint logic programming, fuzzy answer set semantics, fuzzy formal concept analysis, logical approaches to qualitative reasoning, and algebraic structures for computer science. He has authored or coauthored more than 120 papers in scientific journals and proceedings of international conferences. He has co-edited the book Foundations of Reasoning under Uncertainty (Springer-Verlag, 2010), as well as several special issues in scientific journals on mathematical and logical foundations of non-classical reasoning, and on concept lattices and their applications.

**Title:** Adjoints and non-canonical reasoning.

**Abstract:** Computational intelligence must necessarily deal with reasoning mechanisms with are outside the realm of classical logic. Several logical approaches have been developed to reproduce different aspects of non-canonical reasoning. In this talk, we will argue on the usefulness of mathematical tools, in particular the use of adjoints, to formalize some approaches to reason under imprecision, uncertainty, and lack of information.

**David Pearce. **Universidad Politécnica de Madrid, Spain.

**Title:** On Logics for Trust and Honesty.

**Abstract:** In this talk I present two extensions of the well-known modal logics of trust

from (Liau, 2003) formed by adding further axioms. The idea is to interconnect the trust

modality with the individual belief modalities of agents. In the first of our logics

we capture the idea that if an agent i trusts an agent j with respect to a statement

p then i believes that j does not disbelieve p; while in the second logic if i trusts

j concerning p, then i believes that j believes p. In this way we can explicate a

type of trust that is linked to honesty or sincerity. These quite intuitive extensions

of the logic of trust help to solve some unintuitive consequences that arise when

the semantics of trust and belief are independent. As a technical result we establish

the soundness and completeness of these logics with respect to semantics based on neighourhood frames.

**László T. Kóczy. **Faculty of Engineering Sciences. Univ. Széchenyi István, Hungary.

**Title: **Classification and recognition of movement sequences.

**Abstract:** Signal processing and classification as a sub problem of signal processing are well researched areas, but new methods and concepts are presented still today. Handwritten characters are satisfying the definition of signal, if we consider it as a chronologically ordered list of two dimensional coordinate pairs. The recognition of such handwritten characters could be solved with methods known from signal processing and classification.

This work outlines hand-printed (non cursive) character recognition from a signal processing aspect starting with the introduction of various types of handwritten gestures and characters, then a short overview of issues and challenges (e.g. input quality, segmentation, pre-processing etc) of handwriting recognition with common solutions. It is followed by a brief summary of concepts of some uni- and multi-stroke character recognizers to present some examples found in literature.

After that a single- and multi-stroke recognizer family (so called Fuzzy-Based Recognizer or FUBAR) is shown in details based on the Ph.D. dissertation of Alex Tormasi with a high emphasis on the basic concept of the recognition method, the construction of initial fuzzy rule bases with statistical and metaheuristic (bacterial evolutionary algorithm, big bang-big crunch, imperial competitive algorithm, particle swarm optimization) methods. Finally the properties of the metaheuristic methods based on the experiences from the research of FUBAR are also summarized.

**Jozef Pócs** received his Ph.D. degree in 2008 at the Mathematical Institute of Slovak Academy of Sciences.

Since 2007 he has been working as research fellow at the Mathematical Institute of Slovak Academy of Sciences in Kosice. Currently he also works as postdoctoral research fellow at Palacky University Olomouc.

His research interests include abstract algebra and application of algebraic methods to information sciences.

**Title:** Fuzzy concept analysis with preference relations.

**Abstract:** Formal Concept Analysis and its various fuzzy (many-valued) modifications represent methods of data analysis for identifying conceptual structures among data sets. A preference relation, either on objects or attributes, can be seen as an additional information, which should be included to a creation process of a concept lattice. From an algebraic point of view, we discuss some possibilities to include preferences into fuzzy concept lattices. The main emphasis will be on the so-called one-sided concept lattices.

**Sandra Sandri** holds a degree in Computer Science from the Brazilian Federal University of São Carlos, a MSc in Applied Computing from the Brazilian National Institute for Space Research (INPE) and a PhD in Computer Sciences from the University Paul Sabatier (Toulouse-France). She is a senior researcher at INPE and has also worked at the Institute for Artificial Intelligence (Barcelona-Spain). Her work is focused in theoretical and applied developments in Computational Intelligence, particularly in Fuzzy Systems.

**Title:** Current Trends on Computational Intelligence for Space Research in Brazil

**Abstract: ** The main driver of innovation in Space Research in Brazil is INPE, the Brazilian National Institute for Space Research. It is focused in areas such as meteorology and climate change, atmospheric science, space science and space engineering. It also provides services such as weather and climate monitoring, satellite tracking and control, and measuring the amount of forest fires, deforestation, lightnings and air pollution occurring in Brazil. This talk is divided into two parts; first, I will briefly present INPE, followed by some of its Computational Intelligence applications under development. These applications include the use of Neuro-Fuzzy Systems for the prediction of regime change in chaotic systems (related to meteorological phenomena), and the use of evolutionary techniques, such as Genetic Algorithms and Particle Swarms, to learn parameters for radar imagery filters.

** Szilvia Nagy. **Faculty of Engineering Sciences. Univ. Széchenyi István, Hungary

**Title:** Wavelets and their possibilities in computational intelligence.

**Abstract: **Wavelet transform is ideal for image and data processing in many ways. It is easy to calculate wavelet transforms by rather simple convolutional filters. Wavelets are organized in resolution levels, and the wavelet transformed values correspond to spatial (temporal) positions – the higher the resolution level, the denser the grid of these positions. It is easy to find edges, patterns, average behavior and fine-scale behavior in a function or image with wavelets. In many inference systems, the way of processing of data is the key, and wavelets are very good candidates for this purpose.

By omitting wavelet coefficients near zero, the compression of data is also possible. Selecting these non-important coefficients might be done by evolutionary algorithms.

Wavelets are also suitable for solving differential equations. It is usual to solve a problem at various resolutions, however, local refinements are also possible, moreover, prediction of the next finer resolution level coefficients is a recent advancement, where computational intelligence can have significant role.