Artificial intelligence (AI) is often thought of as a futuristic technology and depicted as a rationally thinking human-like robot. The idea of this technology and the endless possibilities it can achieve inspires innovative thinking beyond measure. Aside from what we see on the silver screen, the reality of AI is that it can deliver a wide range of applications. Due to the diverse functionality of these systems, Naval Surface Warfare Center Dahlgren Division (NSWCDD) is committed to exploring and implementing this technology for existing and impending projects.
Part of continuing the commitment for AI technological advancement is providing opportunities for government employees to gain knowledgeable information and situational experience.
Dr. Jeffrey Solka is the Senior Scientific and Technical Manager for Naval Data Sciences at NSWCDD and the Naval Sea Systems Command (NAVSEA) Knowledge Point Champion for digital sciences, quantum sciences and AI/machine learning (ML). He oversees AI/ML systems integration efforts for various naval systems and leads AI/ML science and technology.
Solka said that NSWCDD’s dedicated collaborative efforts with academia, industry sectors, other warfare centers and Department of Defense partners ensure that AI/ML technology is efficiently shared and grants more capability to warfighting systems.
At a recent two-day virtual workshop titled AI for Science, hosted by Nvidia, approximately 30 employees from NSWCDD participated in sessions that provided AI/ML course material and real-life simulations using the technology. The participants learned how to engage with artificial intelligence and machine learning and were able to apply that information to their respective fields.
Although AI is not a new concept, emerging in the early 1990s, the key to developing pioneering products is in the way the technology has increased its computational capability via graphical processing units and cheaper data storage. These capabilities have allowed AI/ML to expand into new areas and applications.
Israel Johnson is a software engineer for the Advanced and Future Maritime Systems Development Branch in the Weapons Control and Integrations Department at NSWCDD and attended the workshop. He serves as a data scientist and provides consultation on data governance. Johnson applies the overall concepts and ideas for other AI/ML projects to become more proficient in gathering data, analyzing the data and determining which form of machine learning applications are best suited for a particular challenge.
As a computer software engineer working in the Software Applications and Systems Security Engineering Branch (Rapid Integration Group), Emily Rounds also attended the workshop. Currently, Rounds works on the Autonomous System Test Capability project, which conducts verification and validation (V&V) of autonomous systems. The project will utilize high performance computing (HPC) systems, which act as an information highway, providing connectivity capabilities and resources to external personnel in remote locations. “Through the use of the HPCs, my team will lead V&V of autonomous systems that will have AI technology embedded into those systems for a wide range of things, including but not limited to object avoidance, object detection and path planning,” said Rounds.
Avory Bryant, a participant in the workshop, is a computer scientist with the Cyber Analytic Advancement Branch for the Strategic and Computing Systems Department. He was already proficient and retained expert knowledge of deep neural networks. Bryant applied the concepts to the multiple projects he is involved in. “One of the problems we’re investigating is techniques for training deep neural networks using synthetic data,” Bryant said, adding that models trained in this way “perform well on real-world data, which is the synthetic-to-real domain adaptation problem.”
With the technological advances many projects and programs are achieving with AI/ML, the need for current and future investments in AI technology is vital for the continued development for systems integration.
As part of the Naval Innovative Science Engineering initiative, internal investments are allocated for the application and advancement of AI technology in such areas as basic research, applied research and prototype development and support technology transition. These AI/ML systems integration efforts endorse newly developed, innovative technology for a multitude of programs that support the warfighter.
“We have a strong interest in applications that are directly important to the warfighter, including artificial data generation, electromagnetic systems data analysis, autonomous systems and submarine launched ballistic missile systems as well as those that are indirectly related, such as supply chain analytics and Human Resources data analytics,” said Solka.
Understanding the way AI systems perform with numerous projects is vital to ensuring that the integration process works effectively and efficiently. Solka stated, “The focus and guarantee for desirable behavior and safety certification needs to be obtained using verification and validation methodologies before AI/ML can be fully integrated into warfare systems.”
Looking forward, the integration and operability for AI technology continue to progress and expand throughout the different NSWCDD departments and across the board. Overcoming challenges with accessibility and availability of hardware and software, endorsing collaborative efforts, providing continuous training opportunities and aligning development for optimum implementation all contribute to successfully supporting the application of AI/ML to enable autonomous unmanned systems.