Title: Chance Discovery: Data-based
Decision for Systems Design
¡¡
Abstract
A chance is an event, which might be rare and uncertain but
is significant for decision making in the dynamic environment. Since
the
birth of the first human, discovering a chance has been required for
his/her survival in the real nature, and nowadays for our survival
in
the social environment. Since we initiated projects and workshops on
Chance Discovery in 2000, we applied this simple and basic idea for
solving
problems in business and design, as well as detecting earthquake
signs.
In this talk, I review the basic concepts in Chance Discovery, and
fundamental methods and tools developed so far. This lecture shows
human's process of chance discovery, compared with data mining.
Based on
this introduction, I introduce scenario maps, the most basic tool
for chance
discovery. It would be shown how scenario maps support user's
creative
decisions. Some species of scenario maps, e.g., KeyGraph and its
successors,
will be presented with some application to real business.
Biography
Yukio Ohsawa is an associate professor in the School of Engineering,
The University of Tokyo. He received Ph.D in Communication and
Information Engineering from The University of Tokyo. He worked also
for School of Engineering Science in Osaka University (research
associate, 1995-1999), Graduate School of Business Sciences in
University of Tsukuba (associate professor, 1999-2005), and Japan
Science and Technology Corporation (JST researcher, 2000-2003). He
initiated the research area of Chance Discovery and series of
international meetings (conference sessions and workshops), e.g.,
the
fall symposium of the American Association of Artificial
Intelligence
(2001). He edited the first book on Chance Discovery published by
Springer Verlag and special issues in international and Japanese
(domestic) journals. Chance discovery is growing: Journal issues has
been published from the international journals, e.g., Journal of
Contingencies and Crisis Management (2001), New Generation Computing
(2003), New Mathematics and Natural Computing (2005), and from
Jounal
on Soft Computing in conjunction with the special issue
on Web Intelligence (2006), etc, and new books are appearing. He is
in
the editorial board the Japanese Socity of AI and the planning board
of New GenerationComputing, and is the TC chair of IEEE-SMC
technical
committee of Information Systems for Design & Marketing.
¡¡
Title: Wind-tunnel Testing for strategy and
market design
¡¡
 |
Professor Edward Tsang
Department of Computer Science
University of Essex, United Kingdom
Email: edward@essex.ac.uk
URL:
http://cswww.essex.ac.uk/CSP
¡¡ |
Abstract
Designers of racing cars and aeroplanes today would typically put
their
designs through wind-tunnel testing. Many business strategies
determine
the success or failure of an enterprise. It is sensible to put these
strategies into action before testing them rigorously. On the other
hand, new markets are being designed all the time; examples include
the
electricity market and traffic markets. The robustness and
efficiency of
these market affect the society. It is inconceivable to put these
markets into operation without testing them scientifically (e.g. to
ask
what-if questions). Agent-based market modeling allows one to
"wind-tunnel test" strategies and market designs. No model is
perfect,
and no amount of testing will guarantee the success of a strategy or
market. However, agent-based models enable us to test complex
behavior
and their interactions. In this talk, I shall define the research
agenda, and explain some of the work that we have done in
wind-tunnel
testing strategies and market designs. Applications include
automated
bargaining and BT's work force scheduling.
Biography
Edward Tsang holds a first degree in Business Administration (major
in
Finance) and a PhD in Computer Science. He is currently a Professor
in
Computer Science at University of Essex. He is also the Deputy
Director
of Centre for Computational Finance and Economic Agents (CCFEA,
http://www.cfea-labs.net), an
interdisciplinary research centre that
applies artificial intelligence methods finance and economics. CCFEA
is
supported by City Associates, members of which include Barclays,
HSBC,
Bank of England, and others City firms.
Edward Tsang has broad interest in artificial intelligence, which
includes heuristic search, computational finance, economic agents,
constraint satisfaction, combinatorial optimisation, scheduling,
evolutionary computation and automated bargaining. He established
and
leads the Constraint Satisfaction and Optimisation Research Group
and
the Computational Finance Research Group at University of Essex.
Edward Tsang chaired the Technical Committee in Computation Finance
and
Economics in IEEE's Computational Intelligence Society in 2004 and
2005.
He is an editor of IEEE Transactions in Evolutionary Computation,
the
Constraints journal, the Journal of Scheduling and The Journal of
Management and Economics. He has served committees and panels to
many
major international conferences and workshops.
Edward Tsang works closely with industry. He had over five years
experience in the commercial sector in Hong Kong. Furthermore, he
has
given consultation to GEC Marconi, BT (was British Telecom), The
Common wealth Secretariat, WestcomZivo and other organizations.
Guided
Local Search, developed in his laboratory, has been embedded in ILOG
DISPATCHER, a commercial product for vehicle routing. A patent on
multi-objective optimization has been obtained with Honda Research
Europe.
Title: Modeling, Control and Filtering of
Multi-Time-Scale Dynamical Systems Based on fuzzy singularly perturbed
models
¡¡
 |
Professor
Fuchun Sun
Dept. of Computer Science and Technology
State
Key Lab of Intelligent Technology and Systems
Tsinghua University
Beijing 100084, P.R. China
Email: fcsun@mail.tsinghua.edu.cn |
Abstract
In this talk, a fuzzy singularly perturbed model is
proposed for the modeling, control and filtering of a class of
multi-time-scale dynamical systems, such as semiconductor
packaging processes, multi-link (joint) flexible manipulators
and large scale flexible space structure. The variant expands
the idea from Takagi-Sugeno type fuzzy system that the dynamics
of a class of distributed parameter systems can be approximated
by softly merging locally linear singularly perturbed model, and
an increased partition of fuzzy space will lead to a more
precise approximation. As a result, the control problem for
multi-time scale dynamical systems can be reduced to that for
fuzzy singularly perturbed system so that plentiful of
theoretical results are hopefully obtained.
Main achievements in this field will be
reported, which are carried out in my research group at National
Lab of Information Science and Technology of Tsinghua University.
These researches include fuzzy singularly perturbed model (FSPM),
its universal approximation, controller synthesis and filtering
for discrete-time and continuous-time multi-time-scale dynamical
systems.
Biography
Dr. Fuchun Sun is a professor in the Department of
Computer Science and Technology, Tsinghua University, Beijing,
China. His research interests include intelligent control,
networked control system and management, neural networks, fuzzy
systems, nonlinear systems and robotics. He has authored or
coauthored two books and over 100 papers which have appeared in
various journals and conference proceedings. Dr. Sun proposed
the new design idea of the neural network (NN) variable
structure control with a sector, developed systematic design
approaches for the discrete-time adaptive control of (rigid and
flexible link) robotic manipulators. Unlike the conventional NN
control using variable structure control, the NN variable
structure control with a sector can guarantee the system
stability outside the network approximation region, and further
improve the system dynamic performance within the network
approximation region. Control approaches proposed have changed
the existing design modes of neuro-adaptive control systems,
i.e. existing neuro-adaptive control is only used in network
approximation region, if outside the network approximation
region, variable structure control is used, while NN variable
structure control is used in the whole control process. Besides,
Dr. Sun also carried out researches on observer-based adaptive
control for nonlinear systems using neural networks, neuro-fuzzy
adaptive control for nonlinear systems based on dynamic
inversion, fuzzy variable structure control, and internetworked
control systems. Now he is worked on the stability synthesis and
control for FSPMs and their application in the modeling, control
and filering for spacerobots with flexible arms and flexible
spacecrafts.
Dr. Sun is the recipient of the excellent Doctoral Dissertation
Prize of China in 2000 and the Choon-Gang Academic Award by
Korea in 2003. He has been a member of the Technical Committee
on Intelligent Control of the IEEE Control System Society since
2006. He serves as a member of the Editorial Board of the
International Journal of Soft Computing £ A Fusion of
Foundations, Methodologies and Applications.
¡¡
Title: Coordination and Self-Organization in
MultiAgent Systems
¡¡
 |
Professor
Gabriel Ciobanu
Formal Methods
Laboratory, Institute of Computer
Science, Romanian Academy,
700505 Iasi, Romania
Email: gabriel@info.uaic.ro
URL: http://thor.info.uaic.ro/~gabriel/ |
Abstract
We consider a system of agents as a layer between services and
the network hosts. We present a self-organizing system of agents
able to offer services with an increase of the speed, and
decrease of the network traffic. A coordination model defines
how the agents interact, and how their interactions can be
controlled; this includes dynamic creation and destruction of
agents, control of communication flows among agents, control of
spatial distribution of agents, as well as synchronization of
actions over time. Various distances and topological relations
are used to explain our self-organizing mechanism. We analyze
the stability of the system, and provide a new result regarding
a weak form of stability depending only on the moment of the
first requests of agents by hosts. Finally we provide a
visualization of the services evolution using self-organizing
maps, and presenting a service clustering structure.
Biography
Dr.Gabriel Ciobanu is a Professor at the Institute of Computer
Science of the Romanian Academy of Sciences. He is also
affiliated to the "A.I. Cuza" University of Iasi, Faculty of
Computer Science. His research interests include Distributed
Systems and Concurrency, Computational Methods in Biology, and
Theory of Programming. He has edited/authored 5 books and over
100 papers on these topics. For his research in these areas, he
received the 2004 Octav Mayer Award and the 2000 Grigore Moisil
Award of the Romanian Academy of Sciences. He was a visiting
academic to Edinburgh University, Paris XI, Tohoku University,
University of Amsterdam (CWI), and National University of
Singapore.
Title: Principle and Methodology of
Computer Games of Chinese Chess
 |
Professor Xinhe Xu
Institute of
Artificial Intelligence and Robotics
Northeastern University, Shenyang, 110004 China
E-mail: xuxinhe@ise.neu.edu.cn |
Abstract
The game problem is very pervasive. As long as some conflict of
interest exists between the two parties in the ¡°chessboard¡±,
game becomes a method which the conflict expresses and gets
solution. Game and countermeasure will become a kind of hot
problems on intelligent system research.
The chess is a kind of intelligent game abstracted from the wars
therefore it is a standard problem in game. The computer game of
chess is always considered as one of the most challenging
topics.
Chess computer game has been developed in full swing, which was
recognized by IBM Deep Blue Supercomputer triumphed over the
reigning World Chess Champion, while computer game on the
Chinese chess which has a long history is puerile. This is the
situation and task we faced.
1. Research sense, current state and application prospect of
computer game.
2. Analysis of the character, classification and complexity of
chess game problems
3. Game problem is typical Discrete Event Dynamic System.
4. Expression of game status and modeling of game procedure
5. Generation and ordering of evolvement operator (move
operator)
6. The way to solve the problem¡ªexpansion and search of the game
tree.
7. Max-min search and multiple heuristic searches
8. The design of opening book and endgame database.
9. Application of system theory, cybernetics, game theory on the
computer game problems.
Biography
Xinhe Xu graduated from Department of Automatic Control,
Northeastern University (NEU) in 1964. Then he has been working
in NEU as yet. He was a Visiting Scholar at Rensselaer
Polytechnic Institute (RPI) from Sept. 1980 to Aug. 1982£¬at
North Carolina State University (NCSU) for 2 months in 1988. He
was promoted as Professor in 1985. Currently he is the Honor
Director of Research Institute of Artificial Intelligence and
Robotics, NEU. He is the Council Member of Chinese Association
of Artificial Intelligence, the Council Senior Member of Chinese
Association of System Simulation. He is also serving FIRA
(Federation of International Robot-soccer Association) as the
Member of Executive Committee. He is the Edition Board Member of
Control and Decision, Journal of System Simulation, Journal of
Northeastern University, Information and Control in China. He
has accomplished more than 40 research projects and published
more than 400 papers of Journal or International Conference.
Prof. Xu's research areas are in control theory and application,
system Simulation, DEDS and HDS, artificial intelligence, fuzzy
control, robotics and robot soccer, computer vision and pattern
recognition, computer game and so on.
Title:
Intelligent System Design for Episodic Memory
on Mobile Daily Life
¡¡
Conventional
way of recording the daily life is to write a diary which can
help the people recall what they did, but it is based on the
human¡¯s limited memory. Recent advances in pervasive and mobile
technology allow us not only to log user¡¯s daily activities
(location, call log, SMS, and proximity information with
Bluetooth) but also to store personal multimedia information
(photo, video and documents), which contains private information
to reveal user¡¯s daily events. Exploiting the information in a
full scale would give us a chance to produce innovative
applications in accordance with personal preference and
contexts.
Automatic diary
generation from user¡¯s log (explicit or implicit) is one of hot
research topics. Sumi et al. designed comic diary to summarize
the user¡¯s conference tour in a comic form. Eagle et al. tried
to develop a diary system based on the log information collected
from cellular phone. Nokia¡¯s LifeBlog provides a way to store
and manage user¡¯s photo, multimedia and SMS in a chronological
manner.
In this talk, I
will present the trend of the research to summarize user¡¯s daily
life based on the information collected from mobile devices, and
report on the design experience of an intelligent system that
collects the available information from Smartphone, extracts the
landmarks in a daily life, and generates the summary of
comic-style diary automatically. Landmark means very relevant or
novel events that are useful to recall a sequence of events, and
the organization must be similar to the human memory structure.
From the logged
information such as GPS, call log, SMS, MP3 play lists, battery
level and photo viewer usage, an ensemble of Bayesian networks
are used to infer the landmark events, and the most probable
candidates are dynamically selected based on the relevance.
Finally, they are converted to the comic cartoons by composing
the comic image components.
