Skip to main content
Article | Mathematics and Computer Science

Peterka works to break data logjam, receives award

Tom Peterka, a computer scientist in Argonne National Laboratory’s Mathematics and Computer Science Division, has received an Early Career Research award from the U.S. Department of Energy’s (DOE) Office of Science.

The Early Career award, now in its eighth year, identifies and supports exceptional, promising young researchers. Peterka will receive $2.5 million over five years for his research on data storage and analysis.

Peterka earned the recognition for his work to redefine scientific data models to be communicated, stored and analyzed more efficiently. As the amount of data grows, scientists’ ability to share and interpret this data is lagging behind.

One cause of the data logjam is the familiar scientific data model that represents raw values at individual positions,” Peterka said. To break out of this spiral, we need to fundamentally rethink how scientific data are represented.”

Peterka and his research group are developing a new, more efficient way to represent data. They can accommodate many types of scientific datasets on high-performance supercomputers, opening the doors to deeper analytical reasoning.

The new model uses continuous, functional approximations and is directly usable in many data analysis and visualization methods. That means for many purposes, we never have to revert back to the original individual data points,” Peterka said.

With the Early Career funds, Peterka will research how to approximate scientific datasets on exascale machines, extremely fast supercomputers that Argonne is helping develop. Peterka and his team are researching how to quantify and reduce the new model’s degree of error and reduce data size. We are also devoting time to applications of the new model, including simulations, experiments, analyses and visualizations,” said Peterka.

Peterka’s new data approach will improve the way scientists manage the large amount of data being produced by faster and faster computers. It is set to transform data-heavy fields such as computational fluid dynamics, astrophysics and even weather prediction.