Keynote Speakers
Janusz Kacprzyk
Janusz
Kacprzyk is a Professor of Computer Science at the Systems Research
Institute, Polish Academy of Sciences. He is an Academician of the
Polish Academy of Sciences (Member of the Polish Academy of
Sciences). Professor Kacprzyk was also visiting professor at many
universities in the USA, United Kingdom, Italy and China. His
research interests include soft computing, fuzzy logic, decision
making, decision analysis and decision support, database querying,
information retrieval, data analysis, data mining, etc. His
publication record is: 5 books, 30 volumes, 300 papers. Professor
Kacprzyk is the President of IFSA (International Fuzzy Systems
Association), Fellow of IEEE and IFSA. He received The 2005 IEEE CIS
Fuzzy Pioneer Award for pioneering works on multistage fuzzy
control, notably fuzzy dynamic programming, and The Sixth Kaufmann
Prize and Gold Medal for pioneering works on the use of fuzzy logic
in economy and management. He is the Editor in Chief of 3 Springer’s
book series, and is on editorial boards of ca. 25 journals.
Professor Kacprzyk has been invited to be a member of the IPC at
more than 200 conferences.
Title: Verbalization in data mining
and decision support: linguistic data summaries, protoforms and
natural language generation
Abstract: We consider some issues related to the so-called data
driven decision support systems (DSSs) which basically try to
support the human decision maker in the process of solving complex
real world decision problems. These DSSs provide tools and
techniques for discovering useful and relevant knowledge – meant as
some relationships between crucial attributes – from sets of data
available to the decision maker, both existing in his/her
organization or company, or obtainable from the outside, e.g., from
the Internet.
A serious problem in the implementation of DSSs is an inherent
discrepancy between the human being and the computer. Notably, for
the human being the only fully natural means of communication and
articulation is natural language, the human being’s cognitive
capabilities are limited and not growing, the human being may be
unreliable, prone to fatigue, etc. These gaps have to be bridged to
successfully implement a DSS.
We consider basically the problem of how to bridge the gap between
the human being and computer with respect to a proper presentation
of results of data mining derived by the system. The form of
presentation should be human consistent. Traditionally, the most
popular approach is here the use of computer graphics to visualize
those results by colorful diagrams, histograms, etc., and this has
been successfully applied in many real world problems.
However, there is another, much less popular form of presentation,
which may be called verbalization, in which the results to be used
by the decision maker are presented in natural language. Clearly,
there are many cases when visualization is impossible, for instance
if a DSS involves activities related to, e.g., eye contact; notable
examples are in transportation or military applications. The
presentation in a voice form, in natural language, does not distract
attention of the decision maker, and is the best here..
We use for this purpose the conceptually and numerically simple
concept of a linguistic data(base) summary the essence of which is
to derive some short sentence(s) in natural language that summarize
the contents (semantics) of even very large data sets that can be
incomprehensible in a raw form to the human being. For instance, in
a personnel database, if we request a linguistic summary on
relations between “age”, “qualifications” and “salary”, it can be
”most older and experienced employees earn much more than the mean
salary”. Clearly, such linguistic summaries must be general, hence
involve imprecise terms that are characteristic to natural language.
We use fuzzy logic to effectively and efficiently represent and
process that imprecision. We show how the derivation of “best”
natural language summaries is related to flexible database querying
so that the process of summarization may be implemented. We show
that the approach is conceptually scalable as linguistic summaries
remain comprehensible and meaningful for data sets of any size due
to the use of natural language.
Then, we consider relations between the linguistic data
summarization and natural language generation (NLG), an important
area of modern information technology, and show that the approach to
linguistic data summarization adopted is closely related to some
types of solutions used in natural language generation which can
make it possible to use more and more effective and efficient tools
and techniques developed in this rapidly developing area.
We present an application to support decision making at a computer
retailer.
Qiang Shen
Professor
Qiang Shen holds the established Chair in Computer Science and is
Director of Research with the Department of Computer Science at
Aberystwyth University, UK. He is also an Honorary Fellow at the
University of Edinburgh, UK. Professor Shen's
research interests include: computational intelligence, fuzzy and
qualitative modelling, reasoning under uncertainty, pattern
recognition, data mining, and real-world applications of these
techniques for decision support (e.g. crime detection, consumer
profiling, systems monitoring and medical diagnosis). He has a total
of 27 years of working experience in these areas.
Professor Shen is currently an Associate Editor of two
premier IEEE Transactions (IEEE Transactions on Systems, Man and
Cybernetics; IEEE Transactions on Fuzzy
Systems). He is also an editorial board member of several other
leading international journals. Professor
Shen was the General Chair of the 16th IEEE International Conference
on Fuzzy Systems (FUZZ-IEEE), held in London, 2007.
Also he has chaired and given keynote lectures at many
other international conferences.
Professor Shen has authored 2 research
monographs, and over 230 peer-reviewed papers (a third of which
appeared in world-class journals), including one which received an
Outstanding Transactions Paper Award from IEEE. He has successfully
supervised over 30 PDRAs and PhDs, including a prestigious British
Computer Society Distinguished Dissertation Award winner.
Title: Intelligent Systems for
Intelligence Data Analysis
Abstract: Failures in the detection of serious crime, including
terrorist activity are not necessarily due to insufficient data, but
rather to difficulties in interpreting the available intelligence.
Automated software systems that model and analyse intelligence data
will provide useful means for the assessment of emerging scenarios
for plausible crimes. This offers assistance in rapidly responding
to the need of devising and deploying preventive measures. This talk
will describe the important challenges which arise in this area, and
which offer great opportunities for the development of intelligent
software systems. It will focus on some recent advances in
computational intelligence in general, and in fuzzy systems in
particular. These advances contribute to the accomplishment of those
tasks essential for intelligence data monitoring (amongst other
applications). The talk will
conclude by identifying some significant potential future
developments.
 |
 |