|
Keynote Speakers
- David Wolfe Corne, Heriot Watt
University, UK
- Hideyuki Takagi , Kyushu University, Japan
- B. Jayaram, Indian Institute of
Technology, Delhi, India
- Dipankar Dasgupta, University of Memphis,
USA
- Mario Koeppen, Kyushu Institute of
Technology, Japan
- Gauri Mittal, University of Guelph, Canada
- Kamala Krithivasan, Indian Institute of Technology, Madras, India
- Hendrik Richter, Leipzig University of Applied Sciences, Germany
---------------------------------------------------------------------------------------------------------------------------------------------
David Wolfe Corne, Heriot Watt
University, UK
Title: Super-Heuristics: Evolving
problem solvers
Abstract:
A small number of independent strands of research since the 1960s have
explored ways to automatically combine individual heuristics to help
solve problem instances, or in some cases to produce new algorithms.
The current buzzword "hyper-heuristics" arises from this activity.
Hyper-heuristics is largely concerned with manipulating the order of
execution of individual heuristics (e.g. such as dispatch rules in
scheduling, or first-fit, best-fit, and so on, in bin-packing) in
search for the best solutions to a given problem instance. Often,
bio-inspired
methods, or reinforcement learning methods, are used to do the
manipulation. A small portion of this overall activity is, in some
contrast, concerned with the development of new algorithms --i.e.
combinations of low-level heuristics that not only solve a given
instance, but are reusable for a class of instances. To distinguish
this type of activity, which I believe is particularly promising, and
far more interesting than using hyper-heuristics for solving single
instances, I call it Super-heuristics. Super-heuristics has
achieved some notable successes so far, and could lead to enormously
more efficient and effective optimization in certain areas of industry.
In this talk I will try to characterize the state of the art in
super-heuristics, and set out several potentially promising ways
forward.
Biography: David is
Director of Research for the School of Mathematics and Computer
Sciences (MACS) at Heriot-Watt University, Edinburgh, UK. MACS is
pre-eminent in Scotland in the Mathematical and Computer Sciences,
compromising part of the Maxwell Institute for Mathematical Sciences (a
joint institute with the University of Edinburgh), and a number of pioneering research centers. He also leads the Intelligent
Systems Laboratory, which maintains a portfolio of substantial
achievements that range through fundamental models of computation,
computational systems biology, computational neuroscience, and advanced
methods for design, optimization and data mining. Following several
years as a research associate in the Department of Artificial
Intelligence at the University of Edinburgh, which led to the extremely
successful problem solving approach now called `hyper-heuristics',
David was a Lecturer (1997) and then Reader (2003) at the University of Reading.
He then took up a Chair in Computer Science at the University of Exeter
in 2004, and moved to his current post in 2006. His continuing research
agenda concerns novel methods for optimization, data mining and machine
learning, as well as strategies for solving large scale problems, with
particular interests in multicriteria problems. He serves on the
editorial boards of many prestigious journals, including Natural
Computation, Theoretical Computer Science (C), the AI Review Journal,
the IEEE/ACM Trans on Computational Biology and Bioinformatics and the
IEEE Trans on Evolutionary Computation. He thinks citation counts are a
very poor and damaging way to evaluate science, but, in case you are
interested, he has a g index of 67 and an H index of 31.
---------------------------------------------------------------------------------------------------------------------------------------------
Hideyuki Takagi , Kyushu
University, Japan
Title: Recent
Topics in IEC Research
Abstract: The first topic of this
talk is to show new types of Interactive Evolutionary Computation (IEC)
application researches. Major IEC applications are optimizing target
systems and creating graphics, images, shapes, sounds, vibrations, and
others. We introduce two new types of IEC applications. The first one
is measuring human characteristics. IEC is an optimization method based
on human subjective evaluation. Likely reverse engineering, we may
measure the evaluation characteristics or mental conditions of an IEC
user by analyzing the outputs from the target system optimized by the
user. The second one is extension of IEC
evaluation. Usually IEC optimizes a target system based on IEC user's
subjective evaluation, i.e. psychological evaluation. We may extend the
evaluation from psychological one to physiological one. We show the
framework of the extended IEC.
The second topic of this talk is to overview researches that try to
reduce IEC user fatigue and show our latest research in this area.
Several approaches have been proposed to reduce IEC user's fatigue;
some of them are improving input/output interface, accelerating EC
search, allowing human intervention into EC search, estimating human
evaluations, and others. Here, we introduce our latest research and
show our view.
Biography: Hideyuki Takagi worked
for the Central Research Laboratories of Panasonic Corporation in 1981
- 1995, and was a visiting researcher at UC Berkeley in 1991 - 1993
hosted by Prof. L. A. Zadeh. He moved to Kyushu Institute of Design in
1995 as an Associate Professor and now works for Kyushu University
since both universities merged in 2003. He had worked on cooperating
neural networks, fuzzy systems, and genetic algorithms and now is
focusing on Humanized Computational Intelligence with interactive
evolutionary computation. Prof. Takagi has worked for IEEE SMC Society
as the Vice-President, the Chair of Technical Committee on Soft
Computing, and an Associate Editor of IEEE Trans. on SMC-B. See Detail
Profile at:
http://www.design/kyushu-u.ac.jp/~takagi
---------------------------------------------------------------------------------------------------------------------------------------------
B. Jayaram, Indian Institute of Technology,
Delhi, India
Title: Gene to
Drug in Silico: A Molecular Approach
Abstract: The
world wide genome sequencing efforts and the concurrent developments in
scientific software implementations on massively parallel computer
architectures grant us the opportunity to dream that drug design could
be undertaken against suitable biomolecular targets to develop
individualized medicine almost in an automated way. Currently however,
without the help of any database, an inspection of a DNA sequence does
not tell us whether it is likely to be a gene and if it is a gene for
messenger RNA, what the likely three dimensional structure of its
protein product is. Also drug design softwares fall short of
expectations even if the structures of drug targets are known.
Addressing the above issues from a physico-chemical perspective, we
have developed a novel semi-empirical physico-chemical model for whole
genome analysis (ChemGenome) based on DNA energetics, an all atom
energy based computational protocol for narrowing down the search space
for locating tertiary structures of small globular proteins
(Bhageerath) and a binding free energy based methodology for active
site directed lead molecule design (Sanjeevini). The ChemGenome could
distinguish genes from non-genes in 331 bacterial genomes and 20
eukaryotic genomes with > 90% accuracy. Bhageerath could
successfully bracket native-like structures to within 3 to 7 Ã… in the
5 lowest energy structures for 50 small alpha helical globular
proteins. The Sanjeevini protocols could sort drugs from non-drugs for
a few drug targets helping in addressing both affinity and specificity
issues in drug design. Progresses recorded in the areas of genome
analysis, protein structure prediction and drug design and the software
tools developed and made freely accessible at www.scfbio-iitd.res.in
together with challenges and promises there of will be presented.
References
1. (a) Dutta,S., Singhal,P., Agrawal,P., Tomer,R., Kritee, Khurana,E.
and Jayaram.B. A Physico-Chemical Model for Analyzing DNA sequences,
2006, Journal of Chemical Information & Modelling, 46(1),
78-85. (b) Poonam Singhal, B. Jayaram, Surjit B. Dixit and David L.
Beveridge. Molecular Dynamics Based Physicochemical Model for Gene
Prediction in Prokaryotic Genomes, 2008, Biophysical Journal, 94,
4173-4183.
2. (a), Narang,P, Bhushan,K., Bose,S. and Jayaram,B. A computational
pathway for bracketing native-like structures for small alpha helical
globular proteins. 2005, Phys. Chem. Chem. Phys., 7, 2364.; (b)
Narang,P, Bhushan,K., Bose, S., Jayaram,B. Protein structure evaluation
using an all atom energy based empirical scoring function, 2006, J.
Biomol. Struct. Dyn., 23, 385-4006. (c) Jayaram et al., Bhageerath,
2006, Nucleic Acid Res., 34, 6195-6204; (d) Jayaram, B.. Decoding the
Design Principles of Amino Acids and the Chemical Logic of Protein
Sequences. Available from Nature Proceedings.
http://hdl.handle.net/10101/npre.2008.2135.1 2008
3. (a) Jain, T and Jayaram, B. An all atom energy based computational
protocol for predicting binding affinities of protein-ligand complexes.
2005, FEBS Letters, 579, 6659; (b) Jain, T and Jayaram, B. A
computational protocol for predicting the binding affinities of zinc
containing metalloprotein-ligand complexes. 2007, Proteins: Structure,
function & Bioinformatics, 67, 1167-1178; (c) Shaikh, S.,
Jayaram. B., A swift all atom energy based computational protocol to
predict DNA-Drug binding affinity and Tm, 2007, J. Med. Chem., 50,
2240-2244; (d) Shaikh, S., Jain. T., Sandhu, G., Latha, N., Jayaram.,
B., A physico-chemical pathway from targets to leads, 2007, Current
Pharmaceutical Design, 13, 3454-3470.
Biography: (i) Ph.D. (1986) City
Univ. of New York, USA.
(ii) Post Doctoral Fellow (1987-88) Columbia University, USA.
(iii) Senior Research Associate (1989-90): Wesleyan University, USA.
(iv) Faculty of the Department of Chemistry, IIT Delhi since 1990.
(Assistant Professor from 1990-95, Associate Professor from 1995-1999
& full Professor Since 2000.
(v) Recipient of Chemical Research Society of India Medal (2000)
(vi) Principal Investigator & Coordinator of Supercomputing
Facility for Bioinformatics & Computational Biology, IIT Delhi
(Since 2001).
(vii) Member of the National Task Force on Bioinformatics of Department
of Biotechnology (since 1998)
(viii) Member of Physical Chemistry Programme Advisory Committee of the
Department of Science & Technology (2004 -2006)
(ix) Member of Organic Chemistry Programme Advisory Committee of the
Department of Science & Technology (2007-2009).
(x) Vice President of Indian Biophysical Society (2006-2008).
(xi) Member of DIT’s Working Group on Bioinformatics (since 2007)
(xii) Head, Department of Chemistry, IIT Delhi (Sept. 2006- Aug. 2009).
(xiii) Coordinator, School of Biological Sciences, IIT Delhi (since
2008).
(xiv) Member of National Committee of IUPAB (2008-2010).
(xv) Member of the Editorial Board of Journal of Molecular Graphics and
Modeling (since 2009).
---------------------------------------------------------------------------------------------------------------------------------------------
Dipankar
Dasgupta, University of Memphis, USA
Abstract: The
biological immune system (BIS) is a highly parallel and distributed
adaptive system. It uses feature extraction, memory, diversity and
associative retrieval to solve recognition and classification tasks. In
particular, it learns to recognize relevant patterns, remember
previously encountered patterns and use combinatorics to construct
pattern detectors efficiently. These remarkable information-processing
abilities of the immune system has inspired an emerging field,
sometimes referred to as the Immunological Computation,
Immuno-computing or Artificial Immune Systems (AIS) that extracts ideas
from BIS to develop computational tools for solving science and
engineering problems.
Over
the last two decades, there has been an increased interest in
immuno-inspired techniques and their applications. In general, some of
such models are intended to describe immunological processes for a
better understanding of the dynamical behavior of the BIS in the
presence of antigens. On the other hand, immunity-based models have
been developed in an attempt to solve wide variety of real-world
problems. In particular, there exist a number of applications in
pattern recognition, fault detection, computer security; also other
applications currently being explored in science and engineering
problem domain. This talk will cover the latest advances in
Immunological approaches and a few real-world applications.
Biography:
Dr. Dipankar Dasgupta is a Professor of Computer Science at the
University of Memphis, Tennessee, USA. His research interests are
broadly in the area of scientific computing, tracking real-world
problems through interdisciplinary cooperation. His areas of special
interests include Artificial Immune Systems, Genetic Algorithms, multi-
agent systems and their applications. He published more than 145 papers
in book chapters, journals, and international conferences. His recently
published a graduate textbook on Immunological Computation, published
by CRC press, September 2008. He also edited two books: one is on
Genetic Algorithms and the other entitled "Artificial Immune Systems
and Their Applications" published by Springer-Verlag, 1999. The book on
Artificial Immune Systems is the first book in the field and widely use
as a reference book. Dr. Dasgupta is a senior member of IEEE, ACM and
regularly serves as panelist, keynote speaker and program committee
member (5-6 per year) in many International Conferences. He first
started (in 1997) organizing special tracks and workshops on Artificial
Immune Systems (AIS) and regularly offered tutorials on the topics at
International Conferences since then. He is an associate editor of
three journals and also is the chair of IEEE Task Force on Artificial
Immune Systems.
------------------------------------------------------------------------------------------------------------------
Mario
Koeppen, Kyushu Institute of Technology, Japan
Title: The (still)
various challenges of Gestalt theory
Abstract: Gestalt
and corresponding Gestalt laws of vision are apparent phenomena of
visual perception that still lack general understanding, despite of
passing more than 100 years after its first mentioning in psychological
literature. In this contribution, we want to promote Gestalt as a kind
of challenge to the naturally and biologically inspired computation
community. Browsing a bulk of existing research literature on the
Gestalt theme, with only a few notable exceptions (like the Helmholtz
principle), there is not much indication for a comprehensive approach
to the understanding of Gestalt, for having explanations about the
means for its application, or for advancement in the provision of
models reflecting the complex interplay of Gestalt laws in a
verificable manner. Said this, currently Gestalt triggers more
questions than answers, and it might slowly become obvious that Gestalt
is more than being just a source of inspiration for new algorithms, or
for stimulating modifications of existing algorithms. It also gets
slowly more clear that the only open issue is not just a lack of
"holistic view" in present science, as it is often stated. It seems
that any further progress in this regard might require a more rigorous
departure from existing computational paradigms and concepts than
expected.
In this talk, the state of research on Gestalt in
engineering sciences, esp. image processing and pattern analysis, will
be critically reviewed, and their strong and weak points will be
evaluated. But moreover, new emerging computational paradigms and
models will be evaluated according to what they might provide to the
understanding of Gestalt. Among these paradigms and models, we can find
the Neural Darwinism, which relates evolutionary concepts to the
processing of the brain, or the recently proposed Cogency
Confabulation, which relates learning with the maximization of a priori
probability, and which is accompanied by a novel neural network
architecture. But going further, Gestalt guides to rather more
fundamental issues of general system modelling, and some relations to
unconventional biological theories like Rosen's MR-systems will be
reflected as well.
Biography: Mario
Koeppen was born in 1964. He studied physics at the Humboldt-University
of Berlin and received his master degree in solid state physics in
1991. Afterwards, he worked as scientific assistant at the Central
Institute for Cybernetics and Information Processing in Berlin and
changed his main research interests to image processing and neural
networks. From 1992 to 2006, he was working with the Fraunhofer
Institute for Production Systems and Design Technology. He continued
his works on the industrial applications of image processing, pattern
recognition, and soft computing, esp. evolutionary computation. During
this period, he achieved the doctoral degree at the Technical
University Berlin with his thesis works: "Development of an intelligent
image processing system by using soft computing" with honors. He has
published around 100 peer-reviewed papers in conference proceedings,
journals and books and was active in the organization of various
conferences as chair or member of the program committee, incl. the WSC
on-line conference series on Soft Computing in Industrial Applications,
and the HIS conference series on Hybrid Intelligent Systems. He is
founding member of the World Federation of Soft Computing, and also
member of the editorial board of the Applied Soft Computing journal,
the Intl. Journal on Hybrid Intelligent Systems and the Intl. Journal
on Computational Intelligence Reserach. In 2006, he became JSPS fellow
at the Kyushu Institute of Technology in Japan, and in 2008 Professor
at the Network Design and Reserach Center (NDRC) of the Kyushu
Institute of Technology, where he is conducting now research in the
fields of multi-objective optimization, digital convergence and
multimodal content management.
------------------------------------------------------------------------------------------------------------------
Gauri
Mittal, University of Guelph, Canada
Title: Advances in
signal and image processing to detect contaminants and microorganisms
in foods for safety and quality
Abstract: Metal,
plastics and glass fragments found in packaged bottles, packages and
containers are big concern for food processors. Using ultrasound to
detect these fragments in the containers poses a challenging task for
signal processing and classification. Various novel signal processing
techniques are developed and discussed. The center frequency pressure
ratio (CFPR), variance, and backscattered amplitude integral (BAI)
methods have approximately the same overall detection rating. Root mean
square (RMS) method has the highest overall detection rating. A hybrid
of these methods has significantly improved object detection rating.
Image processing algorithm is based on the longitudinal (vertical)
tracing of a center frequency component obtained using short-time
Fourier Transform (STFT) in conjunction with a transversal (horizontal)
differentiation of the image pixels. This method has improved ability
to detect small glass fragments inside the bottle. Radial
Basis Function Neural Network (RBF-NN) is used for signal
classification. The output of the RBF layer is determined by the
distance between the input vector and its centroid vector. Successful
classification rate of 95% was achieved using RBF-NN method.
A rapid and cost effective technique for identification and
classification of microorganisms was explored using fluorescence
microscopy and image analysis. After staining the microorganisms with
fluorescent dyes (diamidino-2-phenyl-indole (DAPI) and acridine orange,
AO), images of the microorganisms were captured using a CCD camera
attached to a light microscope. Geometrical, optical and textural
features were extracted from the images using image analysis. From
these parameters, the best identification parameters that could
classify the microorganisms with higher accuracy were selected using a
probabilistic neural network (PNN). PNN was then used to classify the
microorganisms with a 100% accuracy using those identification
parameters.
Biography: Dr. G. S.
MITTAL is a Professor of systems and food engineering at the School of
Engineering, University of Guelph, Guelph, Ontario, Canada. He is an
author of more than 250 refereed journal research papers and 210 other
publications, as well as three books such as ”Computerized Controls in
Food Industry”. He is the recipient of the 1994 John Clark Award
presented by the Canadian Society of Biological Engineering, the 1994
Membro Benemerito Award given by the Colombian Association of Food
Engineers, and International Best Researcher award 2005 & 2007
by Japanese Association of Food Machinery Manufacturers. A registered
professional engineer, professor Mittal received the B.Sc.
(engineering) from India (1969); the M.Sc. (1976) from the University
of Manitoba, Canada, and the Ph.D. (1979) from the Ohio State
University, USA. Mittal presently conducting research in the areas of
system modelling, simulation and optimization; high voltage pulsed
electrical field; sensor and detection systems development; neural
network; image processing; and robotics in food processing. Mittal is
an internationally recognized food engineer with more than 30 years of
professional experience. He was an organizing member of many
international symposiums and conferences; and invited by many
universities and research institutes worldwide for conducting
workshops, lectures and keynote addresses on varied topics. He has
provided consulting services in systems and food engineering areas to
many industries in the last 30 years.
------------------------------------------------------------------------------------------------------------------
Kamala Krithivasan, Indian Institute of Technology, Madras, India
Title: Membrane Computing
Abstract: ‘Natural
Computing’ is an area which is pursued with interest in recent
times. ‘Cellular computing’ is a part of Natural Computing.
One model of cellular computing is membrane computing (P Systems)
initiated by Gh. Paun in 1998, aiming at devising a computing model
inspired by the structure and functioning of living cells. It is
a parallel and distributed model of computing.
In the basic model of P systems, one considers a membrane structure
consisting of several cell-like membranes which are hierarchically
embedded in a main membrane, called the skin membrane. Membranes
with no other membranes embedded in them are called elementary. The
membranes delimit regions, where we place objects. The objects
evolve in nondeterministic maximally parallel manner according to given
evolution rules. The objects can also be described by strings and
then they are processed by string operations. Starting from an
initial configuration, identified by membrane structure and objects in
all regions, and using evolution rules, we get a computation. We
consider a computation complete when it halts, i.e., no further rule
can be applied. The result of the halting computation is defined
on the basis of the objects we obtained in a specified membrane or
expelled from the skin membrane. We may look at membranes to correspond to regions in cells and objects
to chemical compounds and use rules to describe the evolution of those
chemical compounds. In this talk some basic concepts about membrane computing will be
discussed and some variations of the basic model explained. The
talk will also touch upon a new variant ‘Spiking Neural P Systems’.
------------------------------------------------------------------------------------------------------------------
Hendrik Richter, Leipzig University of Applied Sciences, Germany
Title: Evolutionary optimization of dynamic fitness functions
Abstract:
Many real world optimization problems are dynamic, which means that the
fitness function changes with time and that the time scale of these
changes is in the same magnitude as the run-time of the evolutionary
algorithm. It is well-known that evolutionary algorithms are remarkably
successful in solving static optimization problems showing a high
degree of problem difficulty. In recent years it further has been shown
that these problem solving abilities can also be used to tackle dynamic
optimization problems. However, certain modifications in the
algorithmic structure of the evolutionary algorithm are necessary to
make it work in dynamic fitness landscapes. Dynamic optimization means
no longer to find one optimal solution, but to track the movement of
the optimal solution with time. This situation means that the algorithm
must be equipped with some additional schemes which can control,
maintain and occasionally enhance the population's diversity. In the
talk I will discuss recent results in designing evolutionary algorithms
that are fit to perform in dynamic environments. This will cover
approaches dealing with the diversity issue namely random diversity
enhancement schemes, memory schemes and anticipation/prediction
methods. Another topic that will be addressed is the question of how to
find out if the fitness function has changed, also know as the change
detection problem.
Biography:
Dr. Hendrik Richter is Professor of Control Engineering at the School
of Electrical Engineering & Information Technology, Leipzig
University of Applied Sciences (HTWK Leipzig). His research interests
are in two main areas, one is nonlinear dynamics and chaos, the other
is nature-inspired problem solving, particularly evolutionary
computation. Both areas overlap in the topic of designing evolutionary
algorithms for solving dynamic optimization problems. He has serves in
the programme committee of the European Events on Evolutionary
Computation in Stochastic and Dynamic Environments (EvoSTOC) and of the
IEEE Congress on Evolutionary Computation’s Special Session on
Evolutionary Computation in Dynamic and Uncertain Environments (ECiDUE)
for about 5 years.
|
|