SURVICE develops machine-learning solutions across multiple domains to provide innovative capabilities in image analysis, data-driven models and simulation, nonlinear prediction, estimation, and control.
Our customers require trust and accountability in every decision-making process, especially when human lives are at stake. As such, SURVICE specializes in developing ML solutions that are not only accurate but reliable and traceable.
One such example is the active development of graph neural networks (GNNs) for use in physics simulation codes. Such networks are state of the art in physics-informed ML models. Simulation predictions provided by these models are conditioned on, and constrained by, the laws of physics and can be traced back to high-resolution physical interactions.
A highly motivated team, rigorous research and development, and a unique adaptive culture allow SURVICE to provide robust, data-driven solutions for a wide variety of problems. SURVICE actively explores the cutting-edge of ML technology and will continually advance it to adapt to the everchanging technological landscape and growing needs of the U.S. Warfighter.