Scientists at PNNL harnessing advances in deep learning, deep reinforcement learning and generative AI to change how science is conducted and achieve original scientific results and breakthroughs.

Research at PNNL and the University of Texas at El Paso are addressing computational challenges of thinking beyond the list and developing bioagent-agnostic signatures to assess threats.

PNNL computing experts Robert Rallo and Court Corley contribute their knowledge to a recent DOE report on applications of AI to energy, materials, and the power grid.

New research informs the development of foundational models for computational chemistry through hardware-software co-design.

The software democratizes hardware designs by enabling domain scientists to design their own accelerators.

PNNL recently partnered with Amazon Web Services for AWS GameDay, a gamified learning event that challenges participants to use AWS solutions to solve real-world technical problems in a team-based setting.

Researchers investigated how stable nanoparticle suspensions form using facet engineering on hematite nanoparticles, demonstrating that controlling the faceting of nanoparticles can effectively maintain particle dispersity.

In a new approach, quantum flow algorithms are used to simulate many-body systems on quantum computers and effectively describe chemical processes.

PNNL is supporting the Department of Homeland Security Science and Technology Directorate's Chemical Security Analysis Center in improving capabilities to enhance detection and analysis of chemical threats.

Researchers devised a quantitative and predictive understanding of the cloud chemistry of biomass-burning organic gases helping increase the understanding of wildfires.

Reproducible workflows for binning and visualizing nuclear magnetic resonance spectra from environmental samples.

The physics-informed machine learning optimization framework is now PNNL’s most popular open-source repository on GitHub.

PNNL researchers use chemometric analysis to model chemistry in molten salt.

In soil, microbes produce and consume methane. Using a technique called pool dilution, researchers can separate the rate of methane production and consumption from the net rate.

Machine learning methods and data science can uncover hidden patterns in nanoparticle synthesis conditions and property outcomes.

PNNL researchers comment on the “Adversarial Machine Learning” white paper from the National Institute of Standards and Technology.

PNNL is one of the founding partners of the Trillion Parameter Consortium, which was formed to create reliable generative AI models.

A team of scientists at PNNL developed new computational models to predict the behavior of these impurities and reduce the expense and risk related to actinide metal production.

Researchers created a methodology to design custom network topologies that improve the execution time for scientific applications.

PNNL’s DataLife tool analyzes and identifies bottlenecks in scientific workflows.