| I-Light
Applications Workshop
Wednesday, December 4, 2002
8:30 am - 4:30 pm
IUPUI University Place Conference Center, Indianapolis
Abstracts
Real Time Response to Streaming Sensor Data
Beth Plale
Assistant Professor of Computer Science
Indiana University Bloomington
Data-driven applications that accept input from data streams originating
at remote sensors must be supported by low latency middleware that does
not impede rapid response to the occurrence of a complex event in the
environment. Similarly, data streams must be responsive to the needs
of data-driven applications, such as a weather forecasting model, with
regards to synchronizing the data-stream notion of time with the model's
notion of time. Our talk describes the problem and lays out current
and future work we are undertaking in the area.
Cell Modeling/Data Integration through a Web-Based
Information Theory Approach
Peter Ortoleva
Director and Distinguished Professor
Center for Cell and Virus Theory
Department of Chemistry
Indiana University Bloomington
Advanced cell genomic, proteomic, metabolic modeling is being used for
drug discovery, treatment optimization and biotechnical applications.
Data is integrated with the cell reaction-transport model through a
new information theory approach. A variety of data types is thereby
used to calibrate the model, automate model-building, fill in gaps in
the model and assess risk and uncertainty in predictions. Spectroscopy,
microscopy, electrophysiology, NMR and other techniques are thereby
integrated with the modeling activity. To facilitate the accessibility
of our technology to researchers worldwide, and to promote growth of
a unified database of cell physiology, biochemistry, genetics, electrophysiology
and other factors, the entire model/data integration is being implemented
in a Web platform that is a prototype for the next-generation Internet-based
research facility.
Discovery in the Classroom - New Stars in Galaxy
Andromeda
C. A. Pilachowski
Astronomy Department
Indiana University Bloomington
Undergraduate students enrolled in a freshman seminar at Indiana University
Bloomington participated in remote observing on the WIYN 0.9-m telescope
in Arizona using videoconferencing with on-site observers and VNC connection
to the telescope and instrument computers. Students obtained digital
images of the Andromeda Galaxy to discover new, previously unknown,
exploding white dwarfs stars. The course was intended for non-majors
to introduce students to science as a creative activity and a way of
thinking about nature, rather than a body of knowledge. Participating
in original astronomical research, students formulated plans for their
measurements and analysis, carried out their research, and presented
their results to their peers.
Videoconferencing to the telescope allowed the students to participate
in the process of obtaining their data in real time. On-site observers
could describe their work at the telescope and respond to students'
questions. The remote observing was carried out in the Astronomy
Department's WIYN Remote Observing Center, which is designed to look
like a telescope control room, giving further authenticity to the experience.
Use of video conferencing helped the students to understand the nature
of astronomical observing at a professional telescope and the effort
involved to obtain high-quality astronomical data. In addition
to video conferencing, the students could participate in the telescope
operations via VNC connections to the telescope control computers, and
also see the real-time sky camera used for target acquisition.
Data, in the form of multi-megabyte FITS images, were retrieved over
the network so that the students could examine the data immediately
and assess its quality.
These activities were carried out with the assistance of UITS Telecommunications
Division and with the support of an Active Learning Grant from Instructional
Support Services. This assistance is gratefully acknowledged.
Towards an Infrastructure for Large-Scale Information
Analysis, Visualization, Information Retrieval Research & Education
Katy Börner and Javed Mostafa
School of Library and Information Science
School of Informatics
Indiana University
This research in progress aims to develop an advanced infrastructure
for teaching and research in digital libraries, information retrieval,
data mining/analysis, and information visualization. The infrastructure
will comprise an Oracle database of about 15 million records. Bearing
in mind that the ACM portal, CiteSeer, and PubMED currently provide
access to about 361,400, 507,800, and 11 million records respectively,
this database is unique in its size and coverage. Most of the documents
will be available in full text. Software that facilitates a continuous,
automatic update of the database will be in place.
An open source software repository will provide access to services such
as utility programs, data analysis and dimensionality reduction, and
visualization/interaction algorithms. An interchange format based on
metadata standards will be developed to ensure that algorithms can be
combined in multiple ways. All Java-based algorithms can be run in standalone
mode as an applet or application. A standardized software framework
will interlink the network of different databases and services by a
common communication protocol.
All services will run on Solar, IU’s Sun E10000 Research. They
can be contributed or requested via the remote graphical user interface.
The infrastructure will directly support the research of the Information
Processing Laboratory at IUB. Collaborations with researchers at the
School of Informatics at IU Bloomington are currently underway to share
the resources of the infrastructure across campus. In addition, an NSF
funded project on that involves computer scientists, biologists, and
information scientists from both IUPUI and IUB campuses will utilize
this infrastructure.
The infrastructure under development differs from existing resources
by: (1) its uniform, modular, open architecture; (2) its scalability
to handle GB size data sets; (3) its parallel computing infrastructure;
(4) its usage of XML-based, OAI derived communication protocols for
easy integration of new databases and services as well as the serialization
of software packages; (5) detailed documentation of data and code but
also links to related publications; and (6) its online GUI supporting
the request and navigation of diverse information processing jobs for
teaching and research purposes.
Genomes to Grids - Bio Data Distribution for
Grid Computing
Don Gilbert
Indiana University Bloomington
Biologists have discovered many millions of genes and genome
features, now part of the bio-data "library" distributed on
computers around the world. Grid computing methods for finding
and using interesting genome knowledge from this mountain
of data are discussed - their promise and practical concerns for
building usable bioinformatics grids.
|
