For New Mexico LEEP, we are interested in applications of AI/machine learning that harness science-informed approaches applied to commercially-useful problems beyond conventional data mining. We are also interested in advanced computing innovations that can grow into commercial offerings that support the high-performance computing ecosystem.
Successful innovators have access to Los Alamos National Laboratory’s expertise in
applying AI/Machine Learning and data analytics at scale across physics, biology, chemistry, materials, and engineering applications.
In these areas, capabilities include strong expertise in applying artificial intelligence and machine learning (AI/ML) and data analytics at scale along with subject matter expertise across physics, biology, chemistry, materials, and engineering applications. Advanced computing resources may be available to support performance demonstrations including large-scale parallel computing platforms, advanced testbeds for high-performance computing components, quantum and neuromorphic computing.
Specifically in the areas of artificial intelligence and machine learning, Los Alamos develops science-informed approaches to machine learning and data analytics using artificial intelligence. Typical problems involve physical systems with constraints imposed by fundamental properties of physics, chemistry, biology, fluid mechanics, or geologic systems. By applying appropriate underlying physical models to large data sets, the ability to discern features from unstructured data is enhanced. These features can provide predictions of future performance of complex systems, failure modes or early warning of unstable behavior, or allow optimization of operations.
Examples where these techniques have been applied in published scientific work include prediction of earthquake risk from seismic data; optimization of oil/gas production using reservoir production data in hydraulic fractured well systems; object recognition from video and satellite imagery; predicting aging and materials failure modes of complex engineered systems; optimizing materials composition in additive manufacturing applications; automated analysis of hidden flaws in non-destructive evaluation applications.
For advanced computing, Los Alamos National Laboratory is on the frontier of scientific computing for large-scale simulations of physical systems (peta-scale and beyond). We are involved in innovations across the spectrum of advanced computing, from new algorithms for efficient data management, to hardware/software that reduces the energy use of computing systems. Alternate and hybrid computing platforms are of interest that allow optimization of processing speed, data storage, and segregation of computational processes for efficiency. Los Alamos is investigating new hardware and software systems such as hybrid platforms, neuromorphic and quantum computing.
For more information on our capability pillar for Integrating Information, Science, and Technology for Prediction, see https://www.lanl.gov/science-innovation/pillars/ist/index.php