Defense Research Sciences

FY2025 accomplishments - Explored the fundamental questions at the intersection of biological and engineered systems. - Explored software-based methods of assuring safety for language-directed autonomous robots. - Developed applications of emerging mathematical theory to apply to modern technology challenges. - Completed exploration of predictive first principles-based modeling of wafer-bonding of multi-functional materials on arbitrary substrates. - Initiated efforts to expand data science techniques for non-ergodic systems.

FY2025 accomplishments - Initiated development of tabletop dynamical simulators to capture high-resolution data of turbulent phenomenology. - Initiated development of machine language-based algorithms for predictive modeling and automated discovery of universal principles governing turbulence. - Developed systematic methods to discover new numerical algorithms; assess efficacy by rediscovering known numerical algorithms and algorithm structures. - Extended algorithm discovery methods to search for new algorithms and algorithm structures. - Demonstrated optimization methods to systematically discover new algorithms and new algorithm structures, especially new hybrid classical-machine learning algorithm…

FY2025 accomplishments - Created cognitive guarantees that addressed MR vulnerabilities and were expressed in languages suitable for proofs from models. - Built cognitive models for reasoning about users of MR systems with sufficient fidelity relative to human behaviors. - Formalized aspects of MR operations to address multiple cognitive models and major classes of cognitive attacks. - Evaluated model, proof, and guarantee validity using automated reasoning tools.

FY2025 accomplishments - Developed eSBOM formats that provide new types of rich metadata. - Implemented cyber reasoning algorithms and tools that leverage eSBOMs to defend against potential flaws during software development. - Conceptualized approaches for trace back of discovered flaws to derive the triggers and to identify how and where to apply fixes. - Extended testbed for integration, testing, and evaluation of eSBOM formats and trace back capabilities for enterprise-scale applications.

FY2025 accomplishments - Awarded FY 2025 grants for new two-year research efforts across YFA topic areas, establishing a new set of scientific approaches to solve current DoW challenges. - Continued FY 2024 research on new concepts for microsystem, biological, strategic, and tactical technologies; information innovation; and defense sciences by exercising second year funding and by providing continued mentorship by program managers. - Awarded Director's Fellowships for top FY 2023 participants to refine technology further and align to DoW needs.

FY2025 accomplishments - Developed approaches for identifying key attributes of trusted human decision-making in dynamic settings. - Created an alignment framework for developing algorithms that can be tuned to align with specific, trusted humans. - Developed measurement techniques for quantifying the alignment of an algorithmic decision-maker with a human decision-maker.

FY2025 accomplishments - Combined programming language analysis and machine learning techniques to provide an automated C to Rust. - Performed initial assessments of tools in terms of percentages of C functions translated to safe and correct Rust programs, as well as in terms of the performance of the translation tools themselves.

FY2025 accomplishments - Developed methods to decompose scientific, technological, and capability claims into constituent components as well as techniques for automatically assessing component feasibility. - Extended and integrated approaches to address high priority scientific feasibility assessment use cases in collaboration with potential transition partners.

FY2025 accomplishments - Developed methods to incorporate logical knowledge and reasoning into AI systems, enabling assurance in decision support.

FY2025 accomplishments - Explored automated decomposition/abstraction approaches for breaking up a natural language statement into reusable natural language lemmas that, if true, constitute a proof.

FY2025 accomplishments - Engaged frontier AI companies to develop a framework for collaboration and define a mutually beneficial incentive structure for all parties (government, universities, and industry).

FY2025 accomplishments - Compared novel knowledge management tools to foundation models through real-world experimentation in domains of potential military interest. - Transitioned novel knowledge management tools to military-relevant organizations and measured operational effectiveness in stakeholder defined experiments. - Developed human-machine discovery tool for document understanding.

FY2025 accomplishments - Utilized the AI agents' capabilities of continually learning from language and vision to enable human-machine collaborative analysis of image, video, and multimedia documents. - Performed assessments of collaborative human-machine image and language analysis capabilities on challenge problems of interest to potential transition partners in the defense and intelligence communities.

FY2025 accomplishments - Demonstrated a trapped atom gyroscope with single measurement angle rate resolution and scale factor exceeding commercial gyroscopes. - Demonstrated stability and dynamic control over trapped atoms.

FY2025 accomplishments - Benchmarked newly developed materials, devices, and integration and packaging technologies. - Demonstrated components for demonstrating next-generation artificial intelligence, efficient communication, intelligent storage, novel sensing-to-action, and distributed computing architectures. - Identified new research directions and amended new projects to the JUMP 2.0 university research portfolio.

FY2025 accomplishments - Produced initial processor designs and refined processor performance models for enforcing code compartmentalization policies. - Incorporated refined processor performance models in initial implementations of compartmentalization and privilege management tools. - Developed attack campaign test cases for operating systems and conducted initial experiments to measure effectiveness and overhead of compartmentalization and privilege management techniques.

FY2025 accomplishments - Examined current and burgeoning technologies for scalable manufacturing of advanced microsystems. - Investigated potential areas for leveraging commercial microsystems scaling for relevant defense applications.

FY2025 accomplishments - Used state-of-the-art modeling techniques to model the large-scale experiments to identify potential areas for future improvements to turbulent boundary layer flow. - Acquired turbulent boundary layer data at high Reynolds number with smooth and rough wall conditions. - Prepared to conduct experiments with high-energy methods to understand theories behind novel means of forced radioisotope decay. - Completed initial studies of staged accelerator concepts for high energy particle source systems. - Produced and detect electrons and muons at 10 GeV energy levels and beyond. - Developed and initiated preliminary demonstrations of correlated multiphoton sources for sensing, communication, and…

FY2025 accomplishments - Predicted and validated the morphogenic evolution in electrochemical interfaces. - Assessed new experimentally informed modeling techniques to allow for the design of new functional electrochemical interfaces. - Researched and developed mechanisms to optimize microbial biomass powder to meet DoW Nutritional Standards for Operational Rations (NSOR). - Developed safety testing plans for human palatability testing of microbial biomass powder, with input from the U.S. Food and Drug Administration (FDA). - Tested candidate drag reduction technologies on surfaces with complex curvatures. - Demonstrated novel material testing techniques to generate design-relevant properties 10x faster than…