Imprecise probability is understood in a very wide sense. It is used as a generic term to cover all mathematical models which measure chance or uncertainty without sharp numerical probabilities. It includes both qualitative (comparative probability, partial preference orderings, …) and quantitative modes (interval probabilities, belief functions, upper and lower previsions, …). Imprecise probability models are needed in inference problems where the relevant information is scarce, vague or conflicting, and in decision problems where preferences may also be incomplete.The special session is open to contributions on all aspects of imprecise probability. Topics of interest include, but are not limited to:

• models of coherent imprecise assessments
• sets of probability measures, credal sets
• interval-valued probabilities
• non-additive set functions
• qualitative reasoning about uncertainty
• limit laws for imprecise probabilities
• physical models of imprecise probability
• philosophical foundations for imprecise probabilities
• elicitation techniques for imprecise probabilities
• robust statistics
• data mining with imprecise probabilities
• decision making with imprecise probabilities
• algorithms for manipulating imprecise probabilities
• credal classification
• applications.

From environmental issues, to product engineering, to control systems, to medical applications, to customer relationship management, to a myriad of other uses, industry is rapidly embracing uncertainly management and soft computing techniques as a way to understand and adapt to a world that changes faster and faster every day. A special session focused on industrial applications is planned as part of the 2010 conference. Papers focusing on challenges and results from commercial applications are encouraged, as are "lessons learned" from those experiences. This session is open to both industrial practitioners and those working with them in academic settings.

Data streams pose different challenges as compare to the batch sets of data. Yet, most of the existing approaches of data mining, machine learning (including clustering, classification, pattern recognition), process control (including system identification, modelling and control), signal processing (including image/video processing, speech/audio processing), evolutionary robotics etc. are still addressed in off-line mode and the data are assumed to be available in batch form. At the same time, on-line and incremental approaches existed for quite some time in all of these areas, but they all usually assumed the model/cluster structure/classifier/controller to be with a pre-fixed structure, usually selected subjectively by experience.
Many modern day advanced industrial applications, such as robotics, autonomous systems, various defence and security applications, advanced chemical and petro-chemical processes, Internet applications etc. generate with a high rate data streams that require on-line, real-time processing with models and systems with structure that can not necessarily be pre-fixed. While achieving certain results in addressing the problem of managing uncertainties in data streams with the use of eIS a range of problems are still open and require concerted efforts of the research community to be addressed.

The EUSFLAT working group on Machine Learning and Data Mining (DAMI) is organizing a special session as part of IPMU 2010 conference in Dortmund. The objective of the special session is to provide a forum for the discussion of recent advances in the application of fuzzy set methodology and technology to data mining and machine learning problems, both in batch off-line and incremental on-line mode, and to offer an opportunity for researchers and practitioners to identify and discuss recent advances and new promising research directions.
Topics of interest include but are not limited to the application of fuzzy sets, fuzzy systems and related methods in data pre-processing and summarization, handling of incomplete and heterogeneous data, data, text, and web mining methods, pattern recognition, dimensionality reduction and feature selection, interactive and on-line data mining, fuzzy mining of data streams, incremental/adaptive fuzzy clustering, fuzzy classification and fuzzy regression appr, evolving fuzzy systems/classifiers, interpretability and process safety aspects, active and semi-supervised learning, human-machine interaction and the incorporation of background knowledge, mining at multiple levels of abstraction,, database querying and ad-hoc data mining, visualization and presentation of data mining results, postprocessing and summarization of knowledge, parallel and distributed mining algorithms, application cases of fuzzy data mining and machine learning.

Aggregation Operators (AGOPs) play a key role in fuzzy sets theory as fuzzy logic connectives. Many advances in theoretical and applied research are being produced in this field. Traditional aggregation operations such as the weighted average are now acknowledged as particular cases of more general families of aggregation operations, such as Choquet integrals. Triangular norms and conorms, uninorms, symmetric sums, OWA to name a few, are widely used families of AGOP. Along theoretical aspects, an increasing interest to practical applications in now emerging. This requires to face new challenges, regarding computational and domain specific issues.

As Lotfi Zadeh wrote already in 1971, "What we still lack, and lack rather acutely, are methods for dealing with systems which are too complex or too ill-defined to admit of precise analysis. Such systems pervade life sciences, social sciences, philosophy, economics, psychology and many other "soft" fields." And as he said also in an interview in 1994, "I expected people in the social sciences, economics, psychology, philosophy, linguistics, politics, sociology, religion and numerous other areas to pick up on it. It's been somewhat of a mystery to me why even to this day, so few social scientists have discovered how useful it could be.
In 2009, about 15 years after this interview, we have started to establish a broad research area on "Soft Computing in Humanities and Social Sciences" in the European Centre for Soft Computing. For details see: http://www.softcomputing.es/schss2009/en/home.php
In 2010, we will continue with this project and therefore we will have a special session at IPMU 2010, that will point out historical, philosophical, sociological and other aspects of theories of uncertainty, vagueness and fuzziness in the fields of humanities and social sciences.

Digital Earth is a virtual representation of the planet encompassing all its systems and forms, including human societies, conceived to facilitate and promote the use of georeferenced multi-dimensional, multi-scale, multi-temporal information from multiple sources over the Internet.
The development of such a representation requires the management of uncertain information at different levels, such as in spatio-temporal data acquisition and representation, scale transformation, vector to raster conversion and vice versa, heterogeneous multisource spatial data fusion, online geodata discovery, selection and retrieval, etc.
The relevance of the digital earth issue is related not just to sustainable development of the society but also to the challenges and opportunities that it offers to basic and applied research in the ICT field.
Topics of interest include:

• Representation of imprecise and uncertain spatio-temporal data and sensor observations
• Management of spatial data with uncertainty
• Spatial and sensor data fusion under uncertainty
• Qualitative spatio and temporal reasoning
• Classification of spatial data
• Flexible spatial and temporal querying
• Location-based querying and services
• Quality management of information in Web2 spatial applications
• Flexible techniques for geodata discovery and mining
  

The session aims to collect a significant set of papers that shows the progress in the world of Intelligent Databases and Information Systems. An important part of this progress is due to the use of concepts from Fuzzy Logic and Fuzzy Set Theory. Together with the fuzzy relational database model, we can find remarkable works in many other data models such as object-oriented, object-relational, spatio-temporal and multidimensional ones. Moreover, some proposals have appeared with the aim of building soft computing applications in a very wide range of areas relying on the use of these soft extensions of data modelling techniques.
New requirements in current information systems have created new expectations for innovative solutions capable of facing the complexities and quick evolution of current real world problems. This session expects to contribute to meet these expectations by presenting a representative selection of both theoretical and applied works in the mentioned areas.

The use of Soft Computing techniques to solve image processing and analysis problems is an every day more consolidated research line. The aim of this special session is to bring together researchers which solve real-world problems related to video, 3D processing, scene understanding and Image processing using fuzzy logic, neural computing, evolutionary computation, machine learning and probabilistic reasoning. Topics of interest include but are not limited to:

• Video processing, motion detection, segmentation and analysis
• Feature measures and extraction (segmentation, contour detection, ...)
• Image and Video enhancement, coding, compression and transmission
• Image characterization, retrieval and scene understanding
• Image formation, synthesis, registration and quality
• 3D processing, analysis, rendering, physic-based vision
• Applications: Medical imaging, Aerial imaging

Probabilistic graphical models are important tools for knowledge representation and reasoning under uncertainty. In recent times, growing interest has emerged for graphical models whose parameters are only partially specified. Unlike models based on classical Bayesian approaches (e.g., Bayesian nets and random Markov fields), where a precise quantification of each probability measure is required, an 'imprecise' quantification based on sets of (instead of single) probability measures can be also considered. This offers a more realistic modeling of the relations between variables and hence returns more robust results. The counterpart of such an increased expressive power is often a higher computational complexity of the inferences. Yet, important advances have been recently made in the development of efficient algorithms and this opens the door to the application of imprecise probabilistic graphical models to a number of real-world problems.
This special session is open to contributions on any kind of probabilistic graphical models with imprecision, such as, e.g., credal networks, imprecise Markov trees, qualitative networks, possibilistic networks, credal classifiers based on graphical models, interval decision trees, etc..

The session is dedicated to intuitionistic fuzzy sets introduced in 1983 by Atanassov. The papers both on the theoretical aspects of the sets and on their applications are welcome.

The session on uncertainty in network mining will focus on methods from uncertainty and probability theory in network and graph mining. In recent years, network-based methods have consistently shown to provide interesting, nontrivial and fruitful approaches to mining and managing complex datasets of all types. In particular, probabilistic methods have proven to be a principled and successful approach to managing uncertainty. In this special session, new developments and results in the area of uncertainty in network mining will be presented. The session will cover probabilistic network models, uncertainty in social, media and Web 2.0 network modeling, probabilistic graph-based recommender systems, methods for link prediction, the modeling of heterogeneous networks, and probabilistic mining of semantic and ontology-based networks.

Description: Rough set theory, proposed by Zdzislaw Pawlak (1926-2006) in 1982, is a model of approximate reasoning under uncertainty. The main idea is based on representation of complex concepts by their lower and upper approximations. In applications, rough set methodology focuses on approximate representation of knowledge derivable from complex data. It leads to significant results in many areas including, for example, data mining, machine learning, finance, industry, multimedia, medicine, and most recently bioinformatics. The proposed session is a forum to share the most recent results of fundamental research as well as practical efforts in rough sets.

The aim of the Special Session "Fuzzy Numbers and Fuzzy Arithmetic" is to bring together researchers interested in different aspects of fuzzy number theory, like approximations and representations of fuzzy numbers, construction of operators for the exact and approximate fuzzy arithmetic and relevant applications of fuzzy calculus. Presentations focused both on theoretical and applied aspects of the topics mentioned above are welcome.
Topics and scope of the Special Session include (but are not limited to):

• Algorithms for fuzzy arithmetic and fuzzy calculus
• Approximation of fuzzy numbers
• Computation in fuzzy arithmetic
• Decomposition of fuzzy numbers
• Distances between fuzzy numbers
• Fuzzy arithmetic
• Fuzzy intervals
• Fuzzy numbers in soft computing
• Generalized fuzzy numbers
• Operations on fuzzy numbers
• Ranking fuzzy numbers
• Representations of fuzzy numbers

http://www.ibspan.waw.pl/~pgrzeg/ipmu2010.htm

The theory of belief functions is considered as a useful theory for representing and managing uncertain knowledge. This theory was introduced by Shafer as a model to represent quantified beliefs. It is also called Dempster-Shafer theory or evidence theory. This approach has been increasingly applied in several fields like artificial intelligence, machine learning, datamining, graphical models, information fusion, web semantics, etc. This session is dedicated to papers presenting current developments, and results of various aspects of the use of the belief function theory to manage knowledge. So, original papers with theoretical and/or practical contributions are solicited. This session includes (but is not limited to) the following topics related to the use of belief function theory for the management of knowledge namely Machine learning, Datamining, Graphical models, Web semantics, Information fusion, Business decision, etc.

This special session addresses the issues of modelling and reasoning with uncertainty arising from imperfect information. The imperfections are of various kind as occurring in the real world. For instance, some information needed may be imprecise, uncertain, incorrect, or not available, or may origin from sources that cannot be fully trusted (e.g. open sources like the web). Further, information from one source may be fully or partly conflicting with information from another source or with existing knowledge. Submissions may consider issues in modelling and managing such uncertainty for situation awareness and/or decision support, including related end-user issues, such as visualization and traceability. Submissions may focus on theory or consider an application in a concrete domain.

For the modelling of uncertainty, probability theory and many other numerical methods including e.g. possibilistic and ranking approaches provide a rich fundus of concepts and tools, allowing for a fine-grained representation of uncertainty. On the other hand, in the field of nonmonotonic reasoning, powerful qualitative, usually strictly logic-based approaches have been developed for the representation of and reasoning with uncertain, defeasible knowledge.
The aim of this special session is to bring together researchers working in both fields, with a focus on joining quantitative, numerical and qualitative, logic-based approaches to modelling uncertainty; of special interest are concepts, methods, tools or applications that try to employ insights or achievements from both sides.
We welcome papers on the following and any related topics:

• representation of uncertain knowledge
• foundations of uncertainty and NMR
• axiomatic properties of uncertain reasoning
• logics for modelling uncertainty
• probabilistic reasoning
• possibilistic approaches
• default reasoning and ranking functions
• logic of conditionals
• probabilistic graphical models
• belief change and uncertainty
• argumentation and uncertainty
• fusion of uncertain knowledge sources
• agent modelling and uncertainty
• uncertainty in application domains

Hundreds of books were published recently on Knowledge Management (KM) and Information Management (IM), worldwide. Unfortunately there is no precise definition of what these concepts mean. Giving the gist of what we learned from dozens of publications: KM is managing knowledge and IM is managing information; even valid definitions of knowledge and information are missing. In this special session we want to find answers to the following questions:

• What makes a message information and what is knowledge
• Does Shannon-Information fit to the concept of KM in economics
• Is expedient information measurable for economical models
• How can such latest findings influence modern IM and KM.

The subject of collective choices was revived in the middle of twentieth century by Kenneth Arrow who was concerned with the difficulties of group decisions and the inconsistencies they can generate, leading to the well- known Impossibility Theorem. The difficulties highlighted by the Theorem and its extensions have stimulated the development of alternative approaches and a number of authors, over the last two decades of the century, have extended the Arrow's approach in order to encompass fuzziness in individual and group preferences, aiming at escaping from impossibility through the idea of fuzzy rationality. It was in this context that the traditional approach to modelling of consensus as a strict and unanimous agreement has been extended assuming that the human perception of consensus is much softer, and decision makers in every day life are willing to accept that consensus has been reached when most of them agree on the preferences associate to the most relevant alternatives. Some authors addressed the problem of modelling consensus in collective choices introducing the use of linguistic preferences to define linguistic consensus degrees.
Topics of interest include:

• Fuzzy preference relations in group decisions
• Aggregation of imprecise individual preferences
• Consensus reaching processes under vagueness
• 'Soft' consensus measures
• Consensus and proximity measures
• Linguistic opinions in group decision making
• Fuzzy majority rules

Evolutionary algorithms are search methods that mimic processes that occur in natural evolution in order to solve complex optimization problems. A population of solutions evolves by means of selection, reproduction and recombination over time. Evolutionary algorithms find solutions in a robust and efficient manner in optimization problems such as structural optimization, hardware in the loop optimization, game strategies, combinatorial optimization, scheduling, artificial life, machine learning, aerodynamic design, control system design and robotics.
The objective of the special session is to bring together researchers to report and discuss recent developments and trends in evolutionary algorithms including but not limited to these topics

• genetic algorithms
• evolution strategies
• evolutionary computation in games
• design of experiments
• multi objective evolutionary algorithms
• theoretical foundations of evolutionary algorithms
• model assisted evolutionary algorithms
• estimation of distribution methods
• evolutionary robotics
• genetic fuzzy systems
• evolutionary structural optimization

Nowadays privacy and security systems have to deal with complex and heterogeneous information, highly networked relations, mobile components, or erratic information and persons. The use of artificial intelligence techniques is emerging in the areas of privacy and security, as means to cope with this issues. This special session attempts to show the relevance of uncertainty and related tools in artificial intelligence in the fields of security and privacy: from areas where the use of artificial intelligence techniques has been traditionally accepted such as intrusion detection systems, and network traffic analysis, to more novel areas such as access control, cryptography, authorization, or data privacy. These last fields have been traditionally associated with absolute and strict approaches, but recently their complexity, the consideration of human beings in the process, or the need to introduce more flexibility in favor of usability, are boosting researchers to consider more soft approaches. Techniques from fuzzy logic, to clustering, expert systems, constraint satisfaction, or pattern matching, are being interestingly applied to several fields of security and privacy.

Since the pioneer works of G. Choquet and M. Sugeno, fuzzy measures and integrals have encountered increasing interest and investigations in areas such as fuzzy set theory, measure theory, decision theory, game theory and artificial intelligence. The Choquet and Sugeno integrals are the best-known examples of fuzzy integrals. The domains of application of the fuzzy measures and integrals are very wide since they cover among many others, multi-criteria decision aiding, decision under uncertainty, data fusion, and classification. The main asset of fuzzy measures is to capture the non-additivity of the various dimensions that are aggregated. This allows representing the interaction between the criteria, the synergy among the players, the redundancy among the features, and also complex decision behaviours such as a vetoes or risk aversion.
This session will address both theoretical and applied aspects of the fuzzy measures and integrals. The topics of this special session include, but are not limited to:

• applications of the fuzzy integrals
• methods for the identification of the fuzzy measure
• Mobius transform, Shapley value of fuzzy measures
• fuzzy integrals on combinatorial structures
• generalized fuzzy integrals
• sub-families of fuzzy measures (e.g. k-additive) or generalized fuzzy measures (e.g. bi-capacities)

This session is dedicated the concept, design and development of the sensing web project (SWEB). SWEB aims at using the existing multiple sensors (e.g. security cameras, traffic cameras, etc.) towards communicating the state of the environment atr a given moment.

Soft Computig is already in position when there are many sophisticated theoretical methods developed in it and there have been also realized various kinds of applications. Not all of them, however, have been supported well by the theory. This is also one of the reasons for partly sceptic relation to SC. It is time to revise some seemingly notoriously known concepts and methods and to improve the applications from the point of view of better understanding the theory. The goal of this session is to collect papers presenting theoretically well justified application(s).

Information fusion approaches are more and more used in industrial applications in which there is a real need to take into account several kinds of information simultaneously. Fusion systems become complex because they involve all the information treatment chain steps (from the extraction to the decision). They have many parameters and they imply an important computation time. They are also not easy to use and to adjust by the end users.
The objective of the special session is to bring together researchers to report and discuss recent developments in the design of information fusion systems including but not limited to these topics:

• How and which information to extract?
• How to represent and to aggregate the information?
• How to decide that the global result is good?
• How an expert could adjust and use this system?

Representing and reasoning about preferences are important skills in many real-life applications. Preferences are a multi-disciplinary topic of interest to economists, operations researchers, philosophers, psychologists, etc. Different research communities have extensively developed preference theories, most notably AI community and fuzzy community, which often show similarities but also complementarities. This session aims at providing an opportunity to bring together works from both communities. Contributions are expected from both theoretical and application perspectives in fuzzy preference modeling, reasoning with/about preferences, etc.
Topics of interest include:

• (fuzzy) preference modeling
• preference elicitation and representation
• handling preferences in multi-agent systems and group decision making
• preference aggregation and revision
• utility and decision under uncertainty models
• multicriteria decision making
• fuzzy logic based decision making techniques
• Bayesian and other decision making techniques
• database queries
• multiobjective search/optimization
• etc

Assessing, reporting, handling and propagating measurement uncertainty is fundamental in a wide range of engineering and experimental sciences. A uncertain measure can be interpreted as a variable whose exact value is not perfectly know. This lack of knowledge can be induced by a random variation during the measurement process due to the aleatory nature of the physical phenomenon used to achieve the measurement. It can also systematically depend on both the accuracy and precision of the measurement instrument but also rounding error due to sampling and quantization of digital measures. Finally, some lack of knowledge on the measurement process or on the relation between the measured variable and the quantity to be measured can also be part of this uncertainty.
The traditional probabilistic and set-based representations of uncertainty in measurement are not likely to handle all these situations together. Therefore, there is a strong need of new models and methods. This special session aims at collecting papers presenting theoretically or experimentally well justified new approaches. Topics of interest include but are not limited to:

• statistics with imprecise data
• non-additive tools in statistical analysis
• propagation of uncertainties and/ or imprecisions in processes
• new approach for regression
• characterization of the measurement process
• modeling of uncertain data
• modeling of uncertain processes
• use of non-additive approaches in signal processing ? approximate models (i.e. fuzzy transform)
• inversion
• distance measures
• ...