IEEE Computational Intelligence Society

ETTC Task Force on MACHINE LEARNING

Aims and Scope

Machine learning and related research fields, such as data mining and intelligent data analysis, have received a great deal of attention in recent years and, beyond doubt, have established themselves as core elements of intelligent and knowledge-based systems design.  In these fields, a multitude of efficient algorithmic methods for inducing models from data in an automated way and for finding "interesting" patterns and relationships in large data sets have been devised. Such methods exploit the capability of computers to search huge amounts of data in a fast and effective manner. More often than not, however, the data to be analyzed is imprecise and afflicted with uncertainty. In the case of heterogeneous data sources such as text and video, the data might moreover be ambiguous and partly conflicting. Last but not least, patterns, relationships, and models of interest are usually vague and match with the data at best approximately. Thus, in order to make the learning and data mining process more robust or, say, "human-like", methods for searching and learning are needed that are tolerant toward imprecision, uncertainty, and exceptions, have approximate reasoning capabilities and are able to handle partial truth.

Here is where computational intelligence (CI) methods come into play. One of the main concerns of CI, a collection of methodologies whose cornerstones are fuzzy logic, neural networks, and evolutionary algorithms, is to complement classical, "hard" computing techniques with properties of the aforementioned kind. For example, the capability of fuzzy sets to interface quantitative patterns with qualitative knowledge structures expressed in terms of natural language can improve the comprehensibility of extracted patterns considerably, which is a point of major importance in machine learning and data mining. Fuzzy information granulation further allows for trading off accuracy against efficiency and understandability of models. These flexible modeling and knowledge representation capabilities of fuzzy sets are nicely complemented by efficient learning and adaptation techniques developed in the field of neural information processing, as well as robust search and optimization strategies as offered by evolutionary algorithms. In fact, CI solutions are especially effective if the strengths of the different techniques are combined into hybrid systems.

The general goal of the task force is to promote CI-related research in the field of machine learning. Moreover, the task force shall provide a forum for discussions on this topic and a repository for resources on including, e.g., software and benchmark data sets.

Members

• Eyke Hüllermeier, Department of Mathematics and Computer Science, University of Marburg (chair)
• Plamen Angelov, InfoLab21, Lancaster University
• Ulrich Bodenhofer, Institute of Bioinformatics, Johannes Kepler University Linz
• Bernadette Bouchon-Meunier, Université Pierre et Marie Curie, Paris
• Zied Elouedi, Maître de Conférence à l’Institut Supérieur de Gestion de Tunis
• Johannes Fürnkranz, TU Darmstadt
• Fernando Gomide, Department of Computer Engineering and Automation, University of Campinas, Brazil
  
• Larry Hall, Department of Computer Science and Engineering, University of South Florida
• Barbara Hammer, Institute of Computer Science, Clausthal University of Technology
• Francisco Herrera, Department of Computer Science and  Artificial Intelligence, University of Granada, Spain
• Yaochu Jin, Honda Research Institute Europe
• Arthur Kordon, Data Mining and Modelling Group, Corporate Work Processes & Six Sigma Centre, The Dow Chemical Company, Freeport, TX, USA
• Jonathan Lawry, Department of Engineering Mathematics, University Of Bristol, United Kingdom
• Sushmita Mitra, Machine Intelligence Unit, Indian Statistical Institute, India
  
• Tomoharu Nakashima, Department of Computer Science and Intelligent Systems, Osaka Prefecture University
• Witold Pedrycz, Department of Electrical and Computer Engineering, University of Alberta
• Anca Ralescu, Computer Science Department, University of Cincinnati, OH, USA
• Daniel Sánchez, Department of Computer Science and Artificial Intelligence, University of Granada, Spain
• Roman Slowinski, Institute of Computing Science, Poznań University of Technology

• Special session "Fuzzy Machine Learning and Data Mining" at IPMU-2010
• Special session "Computational Intelligence in Machine Learning" at WCCI-2010
• Special session "Managing Uncertainty in Data Streams" at IPMU-2010
• Special session "Learning and Data Mining" at SMPS-2010

• Special session "Machine Learning and Data Mining'' at IFSA-EUSFLAT-2009
• Workshop "Preference Learning'' at ECML/PKDD-2009
• Workshop "Knowledge Discovery, Uncertainty, and Similarity in Case-Based Reasoning" at ICCBR-2009
• Special session "Fuzzy Sets in Computational Biology" at IFSA-EUSFLAT-2009
• Workshop "Evolving and Self-Developing Intelligent Systems" (ESDIS-2009) at SSCI-2009
• Special session "Recent Advances in Evolving Fuzzy Systems" at IFSA-EUSFLAT-2009
• Special session "Applications of CI to benefit society" at FUZZ-IEEE 2009
• 3rd IEEE Workshop on Genetic and Evolving Fuzzy Systems, Witten, Germany, 2008
  
• Special session "Uncertainty in Machine Learning and Data Mining" at IPMU-2008
• Special Session "Computational Intelligence in Games" at IEEE WCCI 2008
• Workshop "Knowledge Discovery, Uncertainty, and Similarity in Case-Based Reasoning" at ECCBR-2008
  
• Workshop "Preference Learning'' at ECML/PKDD-2008

Contact: Eyke Hüllermeier