Tutorials. Entrance free

Soft Computing in Bioinformatics

Tuesday September 6, 2005 – 9:00 to 13:00
	Jim Keller Electrical and Computer Engineering Department University of Missouri-Columbia keller@cecs.missouri.edu
Abstract: This tutorial will go over the basics of creating and using similarities between gene products, i.e., proteins or mRNA. We discuss the various categories of similarity measures beginning with the traditional BLAST scores on gene product sequences and culminating with recent research results in the application of fuzzy set theory to this problem. Clustering, knowledge discovery and applications to microarray experiments will be stressed. The following is a topical outline: Background Genes and Gene Products Sequences Structure Microarrays (expression, hypermethylation) Gene Ontology Other taxonomies Gene Product Similarity Measures Visualization and Clustering Knowledge Discovery Function Summarization About the speaker: James M. Keller received the Ph.D. in Mathematics in 1978. He currently holds the rank of Professor in the Electrical and Computer Engineering Department at the University of Missouri-Columbia. He is also the R. L. Tatum Research Professor in the College of Engineering. His research interests center on computational intelligence: fuzzy set theory and fuzzy logic, neural networks, and evolutionary computation with a focus on problems in computer vision, pattern recognition, and information fusion including bioinformatics, spatial reasoning in robotics, sensor and information analysis in technology for eldercare, and landmine detection. Dr. Keller has coauthored over 225 technical publications. Professor Keller is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE), is a Distinguished Lecturer for the IEEE Computational Intelligence Society (CIS) and for the Association for Computing Machinery (ACM). He is a past President of the North American Fuzzy Information Processing Society (NAFIPS), a past Editor-in-Chief of the IEEE Transactions on Fuzzy Systems, and was the general chair for the 2003 IEEE International Conference on Fuzzy Systems. He is currently the Vice President for Publications for CIS.

 

Soft Computing in Fault Detection and Isolation

Tuesday September 6, 2005 – 9:00 to 13:00
	Józef Korbicz Institute of Control and Computation Engineering University of Zielona Góra J.Korbicz@issi.uz.zgora.pl
	Marcin Witczak Institute of Control and Computation Engineering University of Zielona Góra M.Witczak@issi.uz.zgora.pl
Keywords: fault detection and isolation (FDI), robustness, model uncertainty, neural networks, fuzzy logic, neuro-fuzzy networks, evolutionary algorithms, applications. Abstract: The main objective of this tutorial is to present the 2005 situation of the state-of-the-art concerning the application of soft computing methods to fault diagnosis and supervision systems. Another objective is to show the unsolved and open problems of modern fault diagnosis and supervision that can be solved either with soft computing methods or hybrid systems based on analytical and soft computing methods. The tutorial is divided into five parts. The first part is devoted to the principles of modern fault diagnosis and outlines the state-of-the-art in this important research area with respect to the so-called analytical techniques. Special attention is paid to problems that cannot effectively be solved with such techniques but can be tackled with the help of soft computing methods. The second part is concerned with the design of fault diagnosis schemes with neural networks. In particular, a number of various solutions to modeling problems for fault diagnosis systems are outlined. A special focus is on robustness to model uncertainty, which is very important in practical applications. Various approaches that can be used for tackling this problem are presented, e.g. an experimental design strategy for reducing parametric robustness of a neural model. Hybrid solutions incorporating analytical methods and neural networks are also presented and suitably analyzed. Finally, fault isolation schemes involving neural-network-based classifiers are presented and discussed. The third part is devoted to fuzzy and neuro-fuzzy schemes for FDI. Similarly as for neural networks, attention is focused on modeling problems for fault diagnosis and supervision. Robustness issues with respect to model uncertainty are analyzed as well. Then hybrid solutions such as fuzzy observers or neuro-fuzzy decoupled observers are presented. Finally, fault isolation schemes involving fuzzy- and neuro-fuzzy-based classifiers are presented and carefully discussed. The fourth part deals with evolutionary algorithm-based approaches to the design of fault diagnosis and supervision systems. In particular, various evolutionary schemes that can be utilized to solve modeling problems for FDI are presented, e.g. a genetic-programming-based identification scheme, experimental design determination with evolutionary search with soft selection. Hybrid solutions such as unknown input observer design with genetic programming or robust multi-objective observer synthesis with genetic algorithms are also presented and carefully discussed. Finally, the last part is devoted to case studies and practical implementations of soft computing and hybrid solutions for FDI and supervision problems. In particular, the task of robust fault detection of an industrial valve actuator is tackled with GMDH (Group Method of Data Handling) neural networks as well as with a perceptron neural network obtained with the experimental design strategy. Another study concerns fault diagnosis of an induction motor with a neuro-fuzzy network and genetic-programming-based observers. About the speakers: Józef Korbicz has been a full-rank professor of automatic control at the University of Zielona Góra, Poland, since 1994. He currently heads the Institute of Control and Computation Engineering (ICCE). Born in Poland on 19 March 1951, Józef Korbicz received the M.Sc., Ph.D. and D.Sc. (doctor habilitatis) degrees in automatic control from the Kiev University of Technology, Ukraine in 1975, 1980 and 1986, respectively. After his return to Poland he was appointed an associate professor of automatic control at the University of Zielona Góra. He obtained his professorial title from the Institute of System Research of the Polish Academy of Sciences, Warsaw, in 1993. In 1991 (5 months) he was with the University of Colorado, U.S.A., as an IREX research fellow, and then in 1994 (2 months) with the Universities of Duisburg and Wuppertal, Germany, as a DAAD research fellow. In 1991 he founded the International Journal of Applied Mathematics and Computer Science (AMCS) and up to now he has been the Editor-in-Chief. Moreover, together with prof. J.M. Koscielny, he founded the Polish SAFEPROCESS conferences, the so-called DPP. His current research interests include computational intelligence, fault detection and isolation (FDI) and control theory. The primary aim of his research group is to contribute towards the diagnosis of dynamical systems. His research projects in this field has been sponsored by the State Committee for Scientific Research in Poland and since 1997 by the European Commission: INCO-Copernicus on Integration of quantitative and qualitative fault diagnosis methods within the framework of industrial application, 1997-1999; and 5th FP EU RTN on Development and application of methods for actuator diagnosis in industrial control systems, DAMADICS, 2000-2004. Józef Korbicz has published more than 220 technical papers, 80 of them in international journals. He is a co-author of 8 monographs and text books and a co-editor of 3 books. His last book (co-editor) is entitled Fault Diagnosis. Models, Artificial Intelligence, Applications, Springer-Verlag (2004) (with J.M. Koscielny, Z. Kowalczuk and W. Cholewa). Professor Korbicz is a senior member of IEEE, a member of IFAC TC on SAFEPROCESS, as well as a member of the Automatics and Robotics Committee of the Polish Academy of Sciences in Warsaw. He was a co-chairman of the Programme Committee of the 14th Polish Control Conference, KKA, in Zielona Góra, 2002. Marcin Witczak has been an assistant professor of automatic control and robotics at the Institute of Control and Computation Engineering (ICCE), University of Zielona Góra, Poland, since 2002. Born in Poland on 19 December 1973, Marcin Witczak received the M.Sc. degree in electrical engineering from the University of Zielona Góra (Poland) and the Ph.D. degree in automatic control and robotics from the Wrocław University of Technology (Poland) in 1998 and 2002, respectively. In 2002 (3 months) he was with the University of Hull, United Kingdom, as a research fellow. His current research interests include computational intelligence, fault detection and isolation (FDI), experimental design and control theory. Dr Witczak has taken part in the realization of a number of research projects sponsored by the State Committee for Scientific Research in Poland and since 1998 by the European Commission: INCO-Copernicus on Integration of quantitative and qualitative fault diagnosis methods within the framework of industrial application, 1997-1999; and 5th FP EU RTN on Development and application of methods for actuator diagnosis in industrial control systems, DAMADICS, 2000-2004. Marcin Witczak has published more than 35 papers in international journals and conference proceedings. He is an author of one monograph and four book chapters.

 

Soft Computing for Information Retrieval in the WEB

Tuesday September 6, 2005 – 14:00 to 18:00
	Enrique Herrera-Viedma María J. Martín-Bautista Sergio Guadarrama Alejandro Sobrino José A. Olivas
Abstract: 1. Soft Computing The term SC refers to a family of computing techniques that, when L.A. Zadeh -the father of fuzzy logic- introduced the topic, originally comprised four different partners: fuzzy logic, evolutionary computation, neural networks and probabilistic reasoning. The term SC distinguishes these techniques from hard computing that is considered less flexible and computationally demanding. The key point of the transition from hard to SC is the observation that the computational effort required by conventional computing techniques sometimes not only makes a problem intractable, but is also unnecessary as in many applications precision can be sacrificed in order to accomplish more economical, less complex and more feasible solutions. Imprecision results from our limited capability to resolve detail and encompasses the notions of partial, vague, noisy and incomplete information about the real world. In other words, it becomes not only difficult or even impossible, but also inappropriate to apply hard computing techniques when dealing with situations in which uncertainty and imprecision are involved. The guiding principle of SC is «to exploit the tolerance for imprecision, uncertainty, partial truth, and approximation to achieve tractability, robustness, low solution cost and better rapport with reality». All the methodologies that constitute the realm of SC (the four abovementioned and some others that have been incorporated in the last few years such as rough sets or chaotic computing) are considered complementary as desirable features lacking in one approach are present in another. Hence, the SC framework is put into effect by hybrid systems combining two or more of the constituent technologies with complementary characteristics. 2. Textual Information Retrieval IR may be defined, in general, as the problem of the selection of documentary information from storage in response to search questions provided by a user. IR systems (IRSs) are a kind of information system that deal with data bases composed of information items -documents that usually consist of textual information- and process user queries trying to allow the user to access to relevant information in an appropriate time interval. An IRS is basically constituted by three main components: A documentary base, which stores the documents and the representation of their information contents. It is associated with the indexer module, which automatically generates a representation for each document by extracting the document contents. Textual document representation is typically based on index terms (that can be either single terms or sequences) which are the content identifiers of the documents. A query subsystem, which allows the users to formulate their queries and presents the relevant documents retrieved by the system to them. To do so, it includes a query language that collects the rules to generate legitimate queries and procedures to select the relevant documents. A matching or evaluation mechanism, which evaluates the degree to which the document representations satisfy the requirements expressed in the query, the so called retrieval status value, and retrieves those documents that are judged to be relevant to it. The underlying retrieval model of most of the commercial IRSs is the Boolean one, which is a robust and well formulated model although presents some limitations. For example, it does not consider partial relevance and is not able to rank the retrieved documents by relevance. Due to this fact, some paradigms have been designed to extend this retrieval model and overcome these problems, with the vector space model being the most representative. 3. Web Retrieval Although the textual IR techniques reviewed in the previous subsection are sometimes more than thirty years old, they still constitute the base of modern Web search engines. The popularity of the Web has transformed traditional IRSs into newer and more powerful search tools for locating content on the Internet. However, there are several differences due to the special characteristics of the World Wide Web environment. As Zadeh enunciated in his foreword for F. Crestani and G. Pasi’s edited book on “Soft Computing in Information Retrieval”, the problem of searching the Web has become far more complex that it was in the past mainly due to the increase on the size of the search space by several orders of magnitude and to the multimedia nature of Web documents, being composed of more information kinds than simple plain text. The main existing differences between Web retrieval and traditional IR, highlighting the following ones: The HTML-based nature of Web documents, that make them present a structure defined by the HTML tags. The diversity of Web documents in terms of: i) length, structure, writing style and existence of grammatical and spelling errors; ii) language and domains; and iii) existing information formats, that Web applications have to appropriately deal with. The dynamic nature of many Web pages, that makes their retrieval difficult. The previous aspects clearly show howWeb retrieval have to extend traditional IR in order to deal with the special nature of Web documents. However, this usually makes Web engines focus more on the efficiency of the response than on the retrieval efficacy. Hence, as we shall see in the following section, SC can be a useful tool to build this gap obtaining textual IRSs and Web retrieval engines modelling better the retrieval activity. 4. Soft Computing in Information Retrieval So, what can actually do SC for IR?. Crestani and Pasi gave their view on the answer to this question in the preface of their previously mentioned edited book: “we think that a promising direction to improve IRSs’ effectiveness is to model the subjectivity and partiality intrinsic in the IR process, and to make IRSs adaptative, i.e., able to ‘learn the users concept of relevance’ ”. In a few words, they believe that SC can incorporate a greater flexibility to IRSs and, in view of the characteristics of this research area, it actually seems that this could be the case. On the one hand, the modelling of the subjectiveness and uncertainty existing in the IR activity can be performed by the knowledge representation components of SC such as fuzzy logic, probabilistic reasoning, and rough sets. It is clear that uncertainty and imprecision are involved in the IR activity as, for example, the estimation of the relevance of a document to a user query or the own formulation of a query representing his information needs are pervaded with these characteristics. Concretely, fuzzy logic is a suitable tool to manage the retrieval activity as it is a formal tool designed to deal with imprecision and vagueness and as it facilitates the definition of a superstructure of the Boolean model, so that existing Boolean IRSs can be modified without completely redesigning them. Besides, probabilistic models are powerful and mathematically well formulated techniques to express and handle uncertainty since some decades ago. On the other hand, the IRS adaptativeness mentioned by Crestani and Pasi is related to the machine learning perspective of SC, put into effect by evolutionary algorithms, neural networks and Bayesian networks, among others. These techniques and their hybridizations with IRSs based on the previous knowledge representation approaches can be applied to textual and Web retrieval tasks such as, for example, information extraction and Web mining, inductive query by example and relevance feedback, textual and Web document classification and clustering, and information filtering and recommendation systems. Contents: Problems of the information retrieval and access in the web. Techniques to solve the problems. Classic models of Information Retrieval. Soft Computing and Information Retrieval in the web: Fuzzy Model Applications with Genetic Algorithms Fuzzy Logic tools and problems. Applications and Examples. Speakers: Enrique Herrera-Viedma (1, 2 and 3) María J. Martín-Bautista (4) Sergio Guadarrama (5) Alejandro Sobrino (6) José A. Olivas (6)

 

Soft Data Fusion for the Industrial Application of Computer Vision

Tuesday September 6, 2005 – 14:00 to 18:00
	Aureli Soria aureli.soria-frisch@ieee.org
Abstract: The relevance of information fusion methodologies increases due to the complementary development of computer and sensory technologies. Newly hardware and software facilities allow the inclusion of different information sources in a computer system. Information fusion basically attains the transformation of the information delivered by multiple sources into one representational form. The fused data does not only reflect information that can be extracted from the individual sources but also information not derivable from any of them on its own. Such an information gain characterizes the purpose of information fusion. Operator research in the context of fuzzy systems has generated a fruitful set of aggregation operators, e.g. fuzzy connectives, weighted ranking operators, Ordered Weighted Averaging (OWA) operators, Fuzzy Integrals. So-called fuzzy aggregation operators constitute a flexible alternative to operators traditionally used in information fusion. Among them it is worth pointing out the role of the fuzzy integral. The concept of fuzzy integral is due to Sugeno, who presented in 1974 a mathematical approach within Fuzzy Computing for the simulation of multi-criteria evaluation taking into consideration some cognitive aspects. Sugeno's hypothesis is that the process of multi-criteria integration undertaken by human beings subsumes the linear combination of the different criteria with numerically expressed priorities, i.e. weighted sum strategy. Due to its relationship with cognitive processes and to its positive features as fusion operator, the fuzzy integral is employed in different application fields, where Decision Making and Subjective Evaluation represent the most natural ones. Furthermore fuzzy integrals were used in Computer Vision problems, both on Image Processing and Image Analysis, in a very early stage of research. In this context the fuzzy integral is mainly used because of its mathematical properties as fusion operator, which will be elucidated in the tutorial. In spite of the flexibility, robustness, and interpretability that the fuzzy integral presents when being used as fusion operator, few information fusion applications, especially in Computer Vision, are based on it. This may be due to the complex theoretical background and to the lack of successful implementations of the methodology. Therefore the tutorial brings the fuzzy integral from a mathematical domain to the engineering domain. This goal is achieved in different steps. First, an engineering framework for all fuzzy fusion operators, which is denoted as Soft Data Fusion, is developed. Furthermore different processing frameworks with information fusion, which go beyond the application of the fuzzy integral on its own, are developed. These frameworks are eventually applied for edge detection on color images, for the industrial inspection of high reflective materials, for the processing of document images, the segmentation of color images and for the industrial inspection of end consumer goods. Second the tutorial gives the guidelines underlying the development of different methodologies, which can be employed in the automated parameterization of the fuzzy integral within computational intelligence systems. In this context Soft Computing methodologies present the advantage of being data-driven, what facilitates the implementation of full automated systems for information fusion based on the fuzzy integral. Neurocomputing and Evolutionary Computing are the paradigms selected for the resolution of this problem in the here presented tutorial. Outline Multi-sensory computer vision Computer vision systems Imaging as measuring in different spectral domains Data and multi-sensory fusion Integration vs Fusion Fusion taxonomy in computer vision Application fields of data fusion in computer vision Methodologies for data fusion Soft Computing for data processing and fusion Cognitive inspiration Sugeno’s idea Multi-sensory body Multi-sensory fusion at a cognitive level Multi-sensory fusion at a systemic level Multi-sensory fusion at a neuronal level Application of the fuzzy integral in computer vision Intelligent multi-sensory fusion Soft data fusion Engineering in computer vision with the fuzzy integral Extending the fuzzy integral for image processing Non-automated construction of fuzzy measures Automated construction of fuzzy measures Supervised: genetic algorithms, neural networks, interactive Unsupervised: self-organizing feature maps, statistical analysis Frameworks for image enhancement with soft data fusion Highlights filtering in the industrial inspection of high-reflective materials Color morphology based on the fuzzy integral in the industrial inspection of textiles Frameworks for image transformation with soft data fusion Color edge detection Seal segmentation on tax forms Skin detection on video sequences Frameworks for image analysis with soft data fusion Color image segmentation in market basket recognition Industrial inspection system of collagen plates About the speaker: Aureli Soria-Frisch was born in Barcelona in 1969. He received the 'Enginyer Tčcnic en Telecomunicacions' degree (equivalent BSc) from the University Ramon Llull (Barcelona) in 1992 and the 'Enginyer de Telecomunicació' degree (equivalent MSc) from the Technical University of Catalonia – UPC (Barcelona) in 1995. Since 1996 he is at the Department for Security Technologies of the Fraunhofer IPK (Berlin), where he has participated in several research and industrial projects as research scientist and project leader. He has recently obtained the 'Dr.-Ing.' degree (equivalent PhD) from the Technical University Berlin with a dissertation entitled as "Soft Data Fusion for Computer Vision", which describes the application of the fuzzy integral in different industrial systems. He is author of three journal papers, three book chapters, and several conference papers. He has held different speeches on «Soft Data Fusion» and recently a tutorial at the First Latin-American Summer School on Computational Intelligence (Santiago, Chile). His research interest and expertise are focused on the fields: data and multi-sensory fusion, computational intelligence, soft computing for image processing and analysis, color image processing, texture analysis, and bio-inspired image processing.