neutron by providing data on the internal
structure of materials from the
atomic scale (atomic positions and excitations)
up to the mesoscale (such as
the effects of stress);
Steel production. Steel production
via continuous casting accounts for
an important fraction of global energy
consumption and greenhouse gases
production. Even small improvements
to this process would have profound
societal benefits and save hundreds of
millions of dollars. High-performance
computers are used to improve understanding
of this complex process via
comprehensive computational models,
as well as to apply those models to
find operating conditions to improve
the process; and
Text and data mining. The explosive
growth of research publications has
made finding and tracking relevant
research increasingly difficult. Beyond
the volume of text, principles have
different or similar names across domains.
Text classification, semantic
graph visualization tools, and recommender
systems are increasingly being
used to identify relevant topics and
suggest relevant papers for study.
There are two common themes
across these science and engineering
challenges. The first is an extremely
wide range of temporal and spatial
scales and complex, nonlinear interactions
across multiple biological and
physical processes. These are the most
demanding of computational simulations,
requiring collaborative research
teams, along with the very largest and
most capable computing systems. In
each case, the goal is predictive simulation,
or gleaning insight that tests
theories, identifies subtle interactions,
and guides new research.
The second theme is the enormous
scale and diversity of scientific data
and the unprecedented opportunities
for data assimilation, multidisciplinary
correlation, and statistical
analysis. Whether in biological or
physical sciences, engineering or business,
big data is creating new research
needs and opportunities.