|
|
Track: Digital Media: Content
and Communication
Minitrack: Digital Media at Scale: Confronting the Digital Tsunami
Knowledge Discovery techniques are being overwhelmed by the
need to process greater volumes of disparate data. The perspective
of this problem depends on the constraints of the application.
Some applications have a modest volume of data, but are required
to respond quickly or perform the knowledge discovery on a mobile
device (e.g. PDA) with limited computing and memory. Other
applications have fewer constraints on processing, but must
process enormous volumes of data or high bandwidth data streams.
These applications initially appear to be quite distinct, but
upon closer inspection, they are often examining similar issues.
This minitrack seeks to bring together those working in these
areas to cross-pollinate ideas and advance the state-of-the-art
for this class of problems. Advancements are needed in the ability to address large volumes
of disparate data, streams from distributed locations, short
(often real-time) response or techniques focused on platforms
ranging from mobile to high-performance computing.
Advancements are needed in the ability to address large volumes of
disparate data, streaming in from distributed locations, with
short (often real-time) response. Knowledge Discovery techniques
that address static, modest-sized datasets with unbounded
time-to-answer do not reflect the environment of many end users.
End users in many disciplines, particularly within the
Intelligence Community, cannot adequately confront the digital
tsunami.
The challenges of large-volume data, streaming data, and
reduced response time initially appear to be quite distinct. Upon
closer inspection, they are often to be examining very similar
issues from different perspectives. Foundationally, they have much
in common. Upon closer inspection, they are often found to be
examining very similar issues from different perspectives.
Fundamentally all algorithms trade off space, time, data volume,
and accuracy. If we can solve a problem on a desktop today, we
will expect to be able to solve it on an iPhone or Blackberry
tomorrow. If we solve it with cluster, cloud, or supercomputing
today, we expect to solve it locally, tomorrow. This minitrack
seeks to bring together those working in areas of massive
datasets, streaming data, rapid model construction, across a range
of computational footprints and time constraints.
Both technical and practical presentations are welcomed.
Technical presentations may report either the development of new
methods, an application of existing methods to model, development
of synthetic data sets, or use of massive real data. Practical
presentations may report either on an existing method or heuristic
being applied, an outstanding practical problem which can serve as
a challenge to technical researchers, or similar.
Specifically, we are interested in research and deployment in the
following areas:
-
Massive Data
-
Fusion of
disparate digital media
-
Dealing
with data disparity
-
Fusion of
disparate storage formats for similar data
-
Widely
distributed data
-
Volumes of
data measured in billions of records, terascale through
petascale to yottascale computing
-
Finding
needles in the haystack
-
Understanding overall themes in the haystack
-
Organizing
the haystacks
-
Streaming
Data
-
Response
Times
-
Solutions
tailored to the required response time of the end user
-
Variable response times where the "best rules of action" based
on the information are fluid, changing frequently and
unexpectedly.
-
The
operating rules often require that a good, but sub-optimal
action be selected within a time frame too short to revise and
apply more thorough rules of action.
-
Platform
-
Approaches
-
Novel
approaches that address the issues four points of Massive
Data, Streaming Data, Response Times, and Platform are
solicited.
-
Graph and
Network approaches are of particular interest.
Minitrack Co-chairs:
Mark T. Elmore (Primary Contact)
Oak Ridge National Laboratory
PO Box 2008 MS6364
Oak Ridge TN 37831-6364
Office Phone: 865-241-6372
Departmental Phone: 865-574-4837
Departmental Fax: 865-576-5943
E-mail: ElmoreMT@ornl.gov
Paul Kantor
Rutgers University
4 Huntington St.
New Brunswick NJ. 08901-1071
Office Phone: 732-932-7500 x8216
Departmental Phone: 732-932-7500
Departmental Fax: 732-932-1504
E-mail: Kantor@scils.rutgers.edu
|
