Researchers from Pacific Northwest National Laboratory created and embedded a physics-informed deep neural network that can learn as it processes data.
PNNL-Sequim scientists will spend the next year testing a new technology that could allow the ocean to soak up more carbon dioxide without contributing to ocean acidification.
Randomly constructed neural networks can learn how to represent light interacting with atmospheric aerosols accurately at a low computational cost and improve climate modeling capabilities.
A success story of applying convergence testing to detect and address issues of numerical discretization in nonlinear representations of turbulence and clouds.
Through collaboration with the Department of Homeland Security Soft Target Engineering to Neutralize the Threat Reality Center of Excellence, PNNL is advancing research and development of tools and methodologies to protect crowded places.
A new open-source feature tracking package is now available to facilitate advanced model evaluation, model development efforts, and scientific discovery.
IDREAM research shows that keeping only the most important two- and three-body terms in reactive force fields can decrease computational cost by one order of magnitude, while preserving satisfactory accuracy.
This study demonstrated that a large-scale flooding experiment in coastal Maryland, USA, aiming to understand how freshwater and saltwater floods may alter soil biogeochemical cycles and vegetation in a deciduous coastal forest.
Highly precise and controllable single-atom catalysts are affected by reaction conditions, which can alter the bonding around the atoms and the activity.