- Can geometric algorithms draw networks at least as beautifully as humans are able? How can we make this ability available to thousands of applications?
-
Is it practical to visually explore huge networks?
Can we browse them directly?
Can we run queries on large networks to extract concise
subgraphs that explain interesting relationships?
-
How can we visualize a billion transactions? How should we engineer
systems that provide near-instantaneous access to overviews, mid-level
views, and individual records?
-
How can we extract accurate 3D models from live 2D video streams? What services
can we create from that?
-
Can we use 3D graphics processors and clusters to improve the performance
of algorithms for large-scale data analysis and optimization?
-
Could we make it much easier to analyze large-scale data sets by combining
the best of databases, statistical tools, and visualization in one system?
The Information Visualization Research department contributes
practical techniques for visually exploring and understanding
large, complex data sets.
We are particulaly interested in problems that are technically difficult
due to scale, dimensionality, or complexity of geometric representation.
We place a priority on implementing our work in software components.
Beyond this, we are also experimenting with applications and interfaces for
wall-sized digital displays, which are increasingly important for
collaboration and communication using visualization. We also recently
started a new minilab for work in the integration of 3D graphics with
realtime video, especially in multiple viewpoint camera setups.
Some current projects
include methods for visualizing internet topology data;
rendering clustered networks (such as media recommendations)
as geographic maps;
GPU programming for high performance computing and video processing;
collaboration on a scalable focus+context approach
to displaying and exploring large sets of time series;
a software system that manages over 100,000 servers and routers
with visual querying for data access and exploration.
Information Visualization is part of the
Information, Software and Systems Research Lab (ISSRL)
in AT&T Labs Research. ISSRL, in partnership with the Internet
and Network Systems Research Lab supports the InfoLab,
a multi-disciplinary collaboration to find new ways of turning massive
data sets into useful information, supporting effective
technical and business decision-making. Infolab
researchers include computer scientists, software
and database specialists, statisticians and domain experts.
Our culture emphasizes research
excellence, practical impact, self-motivation
in defining research projects and setting goals,
and opportunities for close collaboration with
experts in related fields of computer science
and mathematics. The lab strives for a balance
between science and applications. Software
skills are highly appreciated.
Beyond the professional environment itself,
one of the main attractions in working here is access to some
of the world's largest networks and services (through AT&T
and its partners) as a source of experimental data and research testbed.
The ISSRL is home to many experts in areas of
computer science, including several
ACM,
IEEE and
AT&T
Fellows.
A sample of local research areas and experts includes:
Statistics Research: Rick Becker,
Bob Bell,
Parni Dasu,
Debby Swayne,
Chris Volinsky,
Simon Urbanek,
Allan Wilks,
Mike Wish
Internet and Information Security Research:
Nick Duffield,
Bala Krishnamurthy,
Carsten Lund,
Walter Willinger,
Jennifer Yates
Algorithms and Optimization Research:
David Applegate,
David Johnson,
Howard Karloff,
Mauricio Resende,
Neil Sloane
Database Research:
Graham Cormode,
Rick Greer,
Marios Hadjieleftheriou,
Ted Johnson,
Flip Korn,
Divesh Srivastava
Software Research:
Mary Fernandez,
Kathleen Fisher,
Bjarne Stroustrup,
Elaine Weyuker,
Pamela Zave
Systems Research:
Robin Chen
Glenn Fowler,
Andrew Hume,
Dave Korn,
Rick Schlichting,
Phong Vo
|
