Information Technology Reference
In-Depth Information
Science at the Speed of Thought
Judith E. Devaney
1
, S.G. Satterfield
1
, J.G. Hagedorn
1
,J.T.Kelso
1
,
A.P. Peskin
1
,W.L.George
1
,T.J.Grin
1
,H.K.Hung
1
,andR.D.Kriz
2
1
National Institute of Standards and Technology, USA
judith.devaney@nist.gov
2
Virginia Tech, USA
rkriz@vt.edu
1
Introduction
Scientific discoveries occur with iterations of theory, experiment, and analysis.
But the methods that scientists use to go about their work are changing [1].
Experiment types are changing. Increasingly, experiment means computa-
tional experiment [2], as computers increase in speed, memory, and parallel
processing capability. Laboratory experiments are becoming parallel as com-
binatorial experiments become more common.
Acquired datasets are changing. Both computer and laboratory experiments
can produce large quantities of data where the time to analyze data can exceed
the time to generate it. Data from experiments can come in surges where the
analysis of each set determines the direction of the next experiments. The data
generated by experiments may also be non-intuitive. For example, nanoscience
is the study of materials whose properties may change greatly as their size is
reduced [3]. Thus analyses may benefit from new ways to examine and interact
with data.
Two factors will accelerate these trends and result in increasing volumes of
data:
-
CPU speedup: as companies strive to keep Moore's law [4, 5] in effect
-
Computer architecture speedup: as all computers benefit from architecture
advances in high end computers.
Figure 1 gives an overview of how these impact problems [6]. These factors
make computers ever more capable and increase the move to computational
experiments and automation.
But a third factor offers a partial solution: graphics speedup. Computer game
enthusiasts are funding a fast pace of development of graphics processing units
(GPUs) [7, 8]. The use of these GPUs in the support of science makes the future
world increasingly computational, visual, and interactive.
We believe that representation and interaction drive discovery, and that
bringing the experiments (computer and laboratory) of science into an interac-
tive, immersive, and collaborative environment provides opportunities for speed
and synergy. Adding traditional data analysis, machine learning, and data min-
ing tools, with multiple representations and interactions can speed up the rate
