|
12th International Conference
on Information Visualisation - IV08 Keynotes
Lectures: |
|
Professor
Georges Grinstein, Director - Institute for Visualization and Perception,
Dr Clive Best,
Joint Research Centre, Arno H.P. Reuser, CEO, Reuser's
Information Services David L. Hicks,
head of Department of Computer Science and Engineering at Aalborg University
of Esbjerg, Denmark Roger Phillips, BSc, MSc, PhD, CITP,
CEng, FBCS, Director and Founder, Vertual Ltd -/www.vertual.co.uk/, Research Professor,
Computer Science, University of Hull, UK |
|
Important Future Research Areas for Information
Visualization
Professor
Georges Grinstein, Director - Institute for
Visualization and Perception,
In looking at current papers in
Information Visualization, one does not see many striking new topics (there are
some). Many papers describe an application of visualization. Most others
describe either an incremental change in a technique to a classic problem or an
improved algorithm reducing computational complexity. There are papers on user
interfaces, on interaction, on usability, on representation, on aesthetics, on
graph drawing algorithms. There are papers on a wide variety of
interdisciplinary topics. There are papers on a new discipline's role in
visualization. Finally there are papers describing something of relevance to
the author but not to the field.
I use to attend the Visualization
Conferences to learn quickly the new major thrusts. But where are these now?
Where are the exciting new topics? What are the exciting problems to be solved?
Are there any problems left? But I still attend – there are useful ideas still
popping up.
Is visualization dead as Bill Lorensen suggested "On the Death of
Visualization" in his presentation at the IEEE Visualization 2005
Conference?
No says I (and he with a caveat).
These are important questions for
a field to identify. Self-introspection is a necessity for a field to continue
to grow.
Now, there have been grand
challenge panels, papers, and pamphlets (in the 70s). I, as a youngster (I am
still young), participated in several of these as far back as 1992 (IEEE
Conference panel on “Grand Challenge Problems in Visualization Software”).
These are valuable. However in my view many of these are driven by the timely
political nature of funding (see for example the excellent "Illuminating
the Path: Research and Development Agenda for Visual Analytics").
Bill Lorensen's
arguments are to focus on more application papers (we, as a field, have and
still do), form alliances with other fields (we have and continue to do so) and
define some grand challenges (though not enough).
So in this talk I will present
five areas which are extremely important for our field and identify key
problems in these five areas. Five areas which can provide for rapid new growth
and which need researchers. And I will identify one in particular which is my
favorite.
These areas are:
1. Measuring Information Visualization
(information, accuracy, uncertainty, insight)
2. High-dimensional Visualization (hundreds and
thousands of variables)
3. Real Time Massive Data Set Visualization
(sensors, networks)
4. Interactive Collaborative Information
Visualization (Web 4.0)
5. Modeling Data Exploration (where is the user
going?)
Georges Grinstein
is Professor of Computer Science at the
He has over 30 years in academia
with extensive private consulting, over 100 research grants, products in use
nationally and internationally, several patents, numerous publications in
journals and conferences, and has been the organizer or chair of national and
international conferences and workshops in Computer Graphics, in Visualization,
and in Data Mining (co-chair IEEE Visualization Conferences, program committee
AAAI conferences in Knowledge Discovery and Databases, co-chair IVBI Symposia,
co-chair IEEE Workshops on the Integration of Databases and Visualization,
co-chair IEEE and AAAI Workshops on the Integration of Data Mining and
Visualization, co-chair ACM workshop on the Psychological and Cognitive Issues
in the Visualization of Data, and co-chair SPIE Visual Data and Exploration and
Analysis Conferences).
He is on the editorial boards of several journals in
Computer Graphics and Data Mining, has been a member of ANSI and ISO, a NATO
Expert, and a technology consultant for various government agencies.
Web Mining
for Open Source Intelligence
Web mining for security involves
the retrieval, extraction and analysis of information from publicly available
sources. Each of these three processes
is the subject of ongoing research resulting in specialised techniques. Today
the largest source of open source information is the Internet. Most newspapers and news agencies have web
sites with live updates on unfolding events, opinions and perspectives on world
events are published. Most governments monitor news reports to feel the pulse
of public opinion, and for early warning and current awareness of emerging
crises. The phenomenal growth in knowledge, data and opinions published on the
Internet requires advanced software tools which allow analysts to cope with the
overflow of information. Malicious use of the Internet has also grown rapidly
particularly on-line fraud, illegal content, virtual stalking, and various
scams. These are all creating major challenges to security and law enforcement
agencies. The alarming increase in the use of the Internet by extremist and Terrorist groups has emerged. The number of terrorist linked websites has
grown from about 15 in 1998 to some 4500 today. These sites use slick
multimedia to distil propaganda whose main purpose is to 1) enthuse and stir up
rebellion in embedded communities 2) instill fear in
the “enemy” and fight psychological warfare. Anonymous communication between
terrorist cells via bulletin boards, chat rooms and email is also prevalent.
The Joint Research Centre has developed
significant experience in Internet content monitoring through its work on media
monitoring (EMM) for the European Commission. EMM forms the core of the
Commissions daily press monitoring service, and has also been adopted by the European
Council Situation Centre for their ODIN system.
JRC has applied core EMM technology to web mining and open source
intelligence for the wider Internet. These techniques use multi-lingual search
techniques to identify potential web resources, followed by the extraction and analysis of the
textual content of web pages. Automatic change detection, the recognition of
places, names and relationships, and further analysis of the resultant large
bodies of text are required. These techniques help analysts to process large
amounts of documents and derive structured data.
This talk will review 3 main
topics:
·
Information
retrieval: Live content monitoring of multilingual news reports. Web scraping
& RSS feed generation, Web Mining and content monitoring
·
Information
Extraction: Topic filtering, Topic Clustering, multilingual named entity
extraction, geocoding and geolocating
text, event extraction, opinion mining.
·
Information
Analysis: Social Network derivation, geospatial indexing and analysis, incident
tracking databases, statistical trend analysis, threat monitoring and
assessment.
Biography of speaker:
Clive
Best has
a PhD in Physics and worked both at CERN and at JET before joining the European
Commission. He has led a group of Web developers since 1994. His team have made
major contributions to Apache and several Web innovations. He has a long
experience of web technologies and currently leads the
Europe Media Monitor (EMM) which monitors world news across the Internet.
The EU depends on EMM for their daily briefings, live updates, health threat
monitoring and crisis alerting. He has contributed to Open GIS standards work,
organised the EOGEO workshop series and
was chairman of a recent NATO advanced study institute on Data mining.
Open Source Intellience
Arno H.P. Reuser, CEO, Reuser's Information
Services, The
Based on many years of experience
in Open Source Intelligence (OSINT) as an information professional and acting head
of the Bureau Open Source Intelligence for the Dutch Defence Intelligence &
Security Service, a model will be presented that forms the basis of Open Source
Intelligence. This Rollercoaster model so called since the work in OSINT
resembles riding a roller coaster: running in a octagram
with the client in the middle, sometimes extremely fast, sometimes not, overall
most exciting and at the end very rewarding -- forms the basis of a process
where the demand side of information and the supply side of information are
matched such that the end user will get the best possible information in the
required format, on the correct subject, within the required time. The place of
OSINT within overall Intelligence will be addressed, as well as the role and
significance of the Internet as an information source for OSINT. The keynote
speech will show why traditional information retrieval systems fail to meet the
demands of today as well as what may be expected of automatic tools.
Especially automatic indexing,
text extraction, entity extraction and meta data generation are essential for
effective information processing in a world that suffers from ˜bytes” overflow.
Arno H.P. Reuser
(a@reuser.biz) was born
and raised in
After graduation, Arno
started his career as scientific librarian at the Central Bureau for Nuclear
Measurements (Euratom) in
This talk discusses some of the
latest research carried out at the Department of Computer Science and
Engineering,
This talk applies several network
centrality measures (and a combination of them) to help identify key players
(important actors) in covert networks.
Most of the models discussed in this talk are implemented in the
investigative data mining toolkit iMiner, which could
be useful for law enforcement agencies that need to analyze terrorist networks
and prioritize their targets.
Structural analysis and
mathematical models for destabilizing terrorist networks are discussed. Three different analysis approaches are
explored:
(i)
Power analysis to uncover a hidden hierarchy within covert networks. A new model of dependence centrality is also
proposed. This centrality measure is
based on shortest paths between pairs of nodes. All models and practical
algorithms are demonstrated.
(ii) Role analysis (such as position role
centrality) to determine key players (for example, gatekeepers) and how their
removal leads to the maximum disruption of a network. The newly introduced measure (i.e., position
role centrality) is more useful in comparison to betweenness
centrality alone because this measure at the same time also can identify
followers.
(iii) Cohesion analysis (such as
cliques, n-cliques, n-clans and k-plexes) to determine familiarity, robustness
and reach ability within subgroups of terrorist networks.
Further, new algorithms for
finding direct and indirect relations in a terrorist network are discussed on
the pattern of graph filtering algorithms.
In addition to exploring
mathematical models, datasets of terrorist attacks that have occurred or were
planned in the past were collected during this study. Data collection is
difficult to do in any network analysis because it is hard to create a complete
network. It is not easy to gain
information on covert networks. These organizations do not provide information
on their members and the government rarely allows researchers to use their
intelligence data. To counter the
information scarcity, a knowledge base has been designed and developed at the
department of Computer Science and Engineering at
David
L. Hicks is an Associate Professor and also serves as Department Head for the
Department of Computer Science and Engineering at
A Vertual Environment Approach for the Training of
Radiotherapy Techniques
Roger
Phillips, BSc, MSc, PhD, CITP, CEng, FBCS, Director and Founder, Vertual Ltd -/www.vertual.co.uk/
Research Professor, Computer Science,
Radiotherapy treatment is the
therapeutic application of radiation to treat cancerous tumours. Over recent
years the techniques and technology have become increasingly more computer based and more complex.
Radiotherapy treatment rooms
provides a less than ideal environment to clinically train staff as the
equipment is very expensive, the equipment needs to be fully utilised to treat
patients and initial training is very stressful for the student in the
treatment room.
We have created a life size
virtual reality treatment room that is suitable for training radiotherapy
concepts and gaining familiarity with the treatment room, its equipment and
delivery techniques.
Virtual reality training is in
many ways superior to training in the real world as it allows the invisible,
e.g. internal anatomy, radiation beams, dose distribution to be visualised in a
patient-centric view of the treatment. Initial studies and use in the field for
over 18 months is proving the worth of this virtual approach to training.
The
Prof. Roger Phillips has
a first class BSc, MSc and PhD degrees in Computer Science from
He
has over 37 years research and development in Computer Science. The first 20
years of this research concerned portability and implementation technology for
compilers, operating systems and software engineering tools needed for their
development.
Since
1992 his research has concentrated on the novel application of computers in
medicine. This has included computer guidance for orthopaedic surgery and
visualization rich methods for training in radiotherapy, interventional
radiology and orthopaedics.
In
the 1990’s he was Head of the Computer Science Department at
He
set up the Hull Immersive Visualization Environment centre in 2002 and was its
Director for its first 6 years of operation.
He
has won numerous awards for his research including the coveted British Computer
Society UK IT awards in 1995 and 2007.
He
is a co-editor of the proceedings of Medicine Meets Vertual
Reality and on the Editorial Board of the Computer Assisted Radiology and
Surgery Journal.
He co-founded
Vertual Ltd in 2007 which is a company that provides
Virtual Reality solutions for the training of radiographers and
radiotherapists.
| TOP |