International Conference on
Machine Learning and Data Mining MLDM´2001
Report about the MLDM 2001
Renaissance Hotel Leipzig/Germany
July 24-27, 2001
Atsushi Imiya, Professor of Media Technology Division, IMIT, Chiba University Japan
One of the origins of Data Mining is pattern recognition in higher dimensional data space. The recent growing applications of data mining are text mining, web mining, and electronic-secretary which automatically classifies thousands of e-mails. In the traditional fields such as computer vision, pattern recognition, and neural network, data mining, knowledge discovery, and machine learning provides new ideas, new understandings and a new framework.
The applications of traditional pattern recognition and artificial intelligence are medical archiving and its applications to computer assisted diagnosis systems and case based system diagnosis. Although these fields have rather long history in artificial intelligence machine learning and data mining and the increasing performance of the computer provide a new paradigm to pattern recognition.
The second interdisciplinary workshop for these fields has been organized at Leipzig by Dr. Petra Perner of IBaI Germany and Prof. Maria Petrou of University of Surrey, England.
The two invited talks in text mining and data mining in clinical application presented up to date subjects. Text mining is an increasing field that is deeply related to the innovation of electronic text and digital libraries. Once texts are transformed to digital forms, it is possible to deal with them as data for pattern recognition. Prof. A. Visa talked about technologies for text mining and showed the challenges for the recognition of similar text part when combining traditional pattern extraction methods with data mining methods. Many rules based on cases are used in clinical diagnosis and case-based system diagnosis. These are old and new fields in artificial intelligence and pattern recognition. Dr. U. Sack - a physician- presented new aspects of clinical applications that can only be discovered when using data mining methods.
24 regular talks are devoted to applications and theory, applications are from geographical information systems, technical diagnosis to medical applications, theories are devoted to neural networks, principal component analysis, and robust statistical methods in pattern recognition and computer vision, such as the extension of EM. New applications of the classification have been presented such as classification of diatoms. In this application computer scientist might help biologist to define new classification criterions. Case-based reasoning and evidence based reasoning become practical applications thanks to the performance of computer. The theoretical consideration of case-base reasoning is one of challenging problems.
There are, now, many conferences named after pattern recognition, computer vision, neural networks, and there are also many workshops devoted to special subjects related these field, for examples, geometry and vision, theoretical aspects or application of machine learning. J. S. Bach established the theory of tonal counterpoint. This music theory is the great present from him to us, which enables us to enjoy great music of 17 century from Leipzig. The fields such as character recognition, sign language understanding, feature extraction from large image sets and robust method for computer vision act as parts of a fuga which produces melodies. On the other hand, data mining and machine learning, act as bass or base of a fuga in pattern recognition.
This is a unique workshop to share the common theoretical and practical background on machine learning and data mining in pattern recognition, computer vision, neural networks, and artificial intelligence. Furthermore, the workshop opens new applications such as text mining to pattern recognition community. The ideas shared at the workshop will bring new ideas in 2003 to the city of J. S. Bach.
The proceedings is available as LNAI 2123, and the publications of the first workshop are also available as LNAI 1715. Selected papers are published in the Pattern Recognition Letters January 2001 as the special issue on MLDM in Pattern Recognition. The selected papers of MLDM 2001 will be appear in the journal Artificial Intelligence in Engineering. Contact address for MLDM 2003 as well as for the new IAPR TC 17 Machine Learning and Pattern Recognition is http://www.ibai-research.de.
MLDM2001 proceedings can be ordered from Springer Verlag: P. Perner (Ed.) Machine Learning and Data Mining in Pattern Recognition, Springer Verlag 2001, lnai 2123, ISBN 3-540-42359