Current Research Opportunities

Reference Code: ICPD-2026-01-HighCapacity

Topic Title: Improving Performance and Feasibility in High-Capacity Power Systems

Keywords: High Energy Density Batteries, Lithium-Ion, Lithium metal, Anodeless batteries, thermal
runaway, battery safety, electrochemical modeling, fuel cells, alternative power

Description: The Intelligence Community (IC) has a pressing need for safe, reliable, and energy-dense power sources to support crucial operations across all mission areas, including security, logistics, and human and technical intelligence. To better protect the interests of the IC’s incorporation of next generation technology with improved performance, reliability, and feasibility, three major initiatives have been identified.

1. Improvements in Lithium Metal Anodes
The IC is persistently interested in improving energy density in batteries with increased temperature performance and cycle life. Lithium metal/anodeless designs with the ability to cycle with no external pressure across extended temperature ranges compared to current lithium-ion technologies would be valuable to the community. Exploring improvements in anodeless lithium metal design to improve safety and cycle life are critical for future implementation of the technology. Proposals of interest may examine electrode coatings, alloys, new electrolyte solvents and salts, or other aspects of cell design that enable desirable operational conditions. Proposal outcomes are anticipated to include descriptions of components and processes to improve cell performance as well as model of baseline and improved cell construction. Validation of design is also desired at the component and pouch cell level
2. Improvements in High-Capacity Cathodes
This topic of interest seeks to develop new cathode materials which surpass the energy density (Wh/L) of state-of-the-art nickel-based cathodes while also demonstrating improvements areas that typically struggle for high-capacity materials. The proposed developed material should strive to show improvements in characteristics such as cycling stability, minimizing overpotentials, low working voltage, rate limitations, mechanical integrity of electrodes, or material stability/safety. Baseline materials of interest should have met or exceeded 1000Wh/L, 50 cycles, 300mAh/g, and >2.0V working voltage in a cell. Materials with industrially relevant packing densities, conductivity, and mechanical stability are also preferred. Proposals of interest should include performance targets and a model for how the proposed material would have a higher energy density than a NMC811 cathode with the same cell design, negative electrode, separator, and operational conditions. T esting and validation at the materials level, coin cell, and pouch cell level should also be included.
3. Logistically available, portable energy sources
While batteries are of interest to the IC, they require recharging infrastructure for continued use. This topic seeks to develop quiet, portable, energy storage systems that can be refueled with globally available, easily procured fuel sources. Diesel/gasoline combustion systems are not of interest. Proposals of interest should address reusable, high energy density systems, (i.e >400Wh/kg), that are recharged/refueled/regenerated without an electrical grid, and preferably from commonly available commercial commodities. Proposals should include an end-to-end system model (including all components such as fuel storage, any reforming required, and power conversion). The desired system should be 1-2 man portable, demonstrate high energy density (>400Wh/kg), supply between 100-1000W, and capable of a minimum of 50 cycles. Validation of fuel source and system design is also strongly desired.

All proposals should discuss the specific barriers to overcome during the research effort, the proposed methods to improve performance, and how the effort is different than past or current research efforts

Example Approaches: 1. Proposals of interest may examine lithium metal electrode coatings, alloys, new electrolyte solvents and salts, or other aspects of cell design that enable desirable operational conditions exceeding curring lithium ion capabilities.

2. The proposed materials should demonstrate baseline performance exceeding current SOA cathode materials, and strive to show improvements in characteristics such as cycling stability, minimizing overpotentials, low working voltage, rate limitations, mechanical integrity of electrodes, or material stability/safety. Materials of interest should meet or exceed 1000Wh/L, 50 cycles, 300mAh/g, and >2.0V working voltage in a cell baseline performance at the beginning of the program.
3. Proposals of interest would demonstrate a 1-2 man portable system exceeding >400Wh/kg, that is recharged/refueled/regenerated without an electrical grid from a commonly available commodity, supply between 100-1000W, and capable of a minimum of 50 cycles.

ORISE Topic Number: ICPD-2026-02-Exothermic

Title: Modeling and Validating Exothermic Reactions in Next Generation Batteries

Keywords: Battery safety, thermal runaway, exothermic reaction, calorimetry, thermal analysis, lithium
metal, dendrites, short circuit, electrolytes, solid state electrolytes

Description: The Intelligence Community (IC) has a pressing need for safe, reliable, and energy-dense power sources to support crucial operations across all mission areas, including security, logistics, and human and technical intelligence. Currently, the IC utilizes various battery chemistries, including lithium-ion, for these essential operations. The need for improved performance in battery technology is critical for next generation technology across the IC. However, while there are many research tracks dedicated to improving battery performance, the change in chemical safety of these higher performances system is not well understood. To better protect the interests of the IC’s incorporation of next generation battery technology and protect people and property with safe deployments and high reliability, this topic area intends to model and validate possible exothermic reaction pathways in next generation battery designs.

Proposals of interest should offer methods to model/understand potential exothermic reaction pathways that could lead to thermal runaway in cells with one or more next generation components including. Components of interested include lithium metal, high silicon content, high nickel cathodes, fluorinated electrolytes, ionic liquid electrolytes, solid-state/semi-solid state electrolytes, phosphate containing materials, additives for improved performance. Development of new materials or electrolytes is not within scope.

Replicating exothermic reactions, including thermal runaway in cells can be difficult to achieve if the reaction pathway is not well characterized and can be affected by a wide number of variables such as state of charge, cycled or calendar age of the cell, or even the cell construction. This topic aims to understand what pathways for energetic cell failure are possible in cells with next generation components that may not readily present under normal operation or safety test conditions, but are fundamentally possible within a certain set of conditions.

Example Approaches: For example, despite many tests demonstrating solid electrolytes have slow discharge rates because of limited ion kinetics, during thermal runaway conditions, the heat generation rate of a LLZO solid electrolyte is capable of reaction rates ten times faster than an equivalent liquid electrolyte (1).Additionally, thermite reactions in LFP cells have also been determined to be thermodynamically possible (2).By understanding the thermodynamics and kinetics of possible exothermic reaction pathways, safer cell designs can be implemented before these processes are discovered by accident in deployed systems.

Proposals of interest may examine reactions under normal cell conditions, as well as degraded or damaged cell conditions where separator damage or breakdown may allow reactions directly between positive and negative cell components. Models of interest should include total heat release, rate of heat release, and reaction byproducts. Validation of reactions of interest is also desirable at the component or cell level, if possible.

Desired Approaches
Proposals to model energetic reactions in next generation batteries should include:
1. Test matrix of components and reaction conditions
2. Model for reactions under standard and damaged cell conditions
3. Plans and description for validation testing of reaction energies and reaction conditionsincluding DSC, TGA, ARC, overheat, overcharge, short circuit
4. Description of validation testing should include whether data is collected from materials, half cell, full cell, etc and any electrode formulation information if applicable.
5. Identification and analysis of components and conditions that may represent high risks for safety/thermal runaway

ORISE Topic Number: ICPD-2026-03-EngSensor

Title: Quantum Engineering for Quantum Sensors

Keywords: Quantum, Quantum Engineering, Quantum Sensors, Atomic Sensors, Machine Learning,
Control Theory, Quantum Control, Signal Processing, Enabling Technology for Quantum
Sensors, Magnetometer, Gyroscope, Accelerometer, Gravimeter, Atomic Clock, Atom
Interferometer, Rydberg sensor, NV Diamond

Description: This topic is about using quantum engineering to make today's quantum sensors easier to build and operate, both in the laboratory and in the field. Quantum sensors are devices that encode a physical quantity into a few quantum states of the system-for example, atomic magnetometers, atom interferometer gravimeters, atomic clocks, Rydberg sensors, NVD magnetometers, and so on. Quantum sensors may optionally utilize non-classical states to increase their performance. As quantum sensors become more sensitive and accurate, a key remaining challenge is to make them more practical outside of the laboratory. They need to be easy to operate, fast to turn on, robust against vibration, thermal changes and electromagnetic noise, small, and low power. Quantum engineering can address these problems by applying both standard and new engineering techniques to quantum devices.

Example Approaches: Example approaches will depend on the maturity of the quantum sensor and its intended application environment. Some interesting directions include (but are not limited to) using machine learning techniques to rapidly optimize sensor performance, using quantum and/or classical control techniques to increase robustness against noise, employing digital signal processing algorithms to increase sensor speed, sensitivity or accuracy, and applying advanced packaging techniques or new materials to reduce sensor size or sensitivity to noise.
Engineering demonstrations may take place in larger, laboratory-scale systems, but the project should not focus on developing new atomic physics techniques and must show how it can be used for smaller, fieldable, sensors. Engineering techniques may also be used to improve the performance of enabling technologies for the quantum sensor, such as lasers, photonic integrated circuits (PICs) or photon detectors, but the proposal should then include the use of these enabling technologies in an actual quantum sensor. Proposals may focus on theory, modeling or algorithms, but must apply these to a quantum sensor in the lab during the first year of the effort.

ORISE Topic Number: ICPD-2026-04-SPECTRA

Title: Systematic Profiling of Enzyme Catalytic Tolerance & Reactivity Atlas (SPECTRA)

Keywords: Biocatalysis, Enzyme Substrate Profiling, Machine Learning, Cheminformatics, LLM agents, Catalytic Boundaries, Substrate Tolerance, Synthetic Biology, Predictive Modeling

Description: Enzyme activity is mapped onto only a fraction of possible substrates, with negative results rarely getting published, leaving the biocatalytic universe largely undocumented and unexplored. This limits our ability to predict how biological catalysts can be repurposed to make high-value chemicals, agrochemicals or toxic scaffolds, as well as to significantly advance the state of the art in AI-informed cheminformatics and biocatalysis.

This opportunity is to systematically profile 8–10 high-impact biocatalysts against 100–200 chemically diverse and obtainable substrates, including pharmaceutical scaffolds and nitrogen-containing heterocycles, to produce a reference dataset that defines substrate-tolerance rules and catalytic boundaries. This experimentally driven atlas will enable high-impact publications and a reliable foundation for downstream predictive modeling.

Example Approaches: Leverage LLM-based agents (multi-step reasoning workflows) to systematically mine the otherwise intractable scattered landscape of enzyme-substrate data across literature, patents, and databases (e.g., Zenodo, BRENDA, ChEMBL) – extracting reported activities, identifying implicit negative results, flagging contradictory findings, and proposing underexplored substrates based on structural gaps in existing data.

Inform and combine this approach with traditional cheminformatics and expert curation to design maximally informative substrate panels for experimental validation. Integrate computational pre-screening methods (e.g., docking, binding prediction) to prioritize substrates based on chemical diversity metrics and predicted tolerance boundaries. Use early-phase results to iteratively refine substrate panels, identifying underexplored chemical space and testing hypotheses about catalytic limits. Systematically capture all data, including negative results, in structured, ML-ready formats to establish a high-quality reference dataset for the biocatalysis community.

Candidate enzyme panel examples:
ω-Transaminase (C. violaceum); Ketoreductase (KRED-P1-A04); Amine Dehydrogenase (Petrotoga mobilis); Cytochrome P450 BM3 (engineered variants); Monoamine Oxidase N (MAO-N); Nitrile Hydratase
(Rhodococcus); Nitrilase (Acidovorax); Aromatic Amino Acid Decarboxylase (AADC); Lipase B (Candida antarctica, CALB); Strictosidine Synthase (Pictet-Spenglerase).

ORISE Topic Number: ICPD-2026-05-SpaceRadiators

Title: High Thermal Flux Space Radiators

Keywords: thermal emission, waste heat, space power, heat exchange, heat rejection, heat transfer, thermal management, thermal control, thermal materials engineering, deployable systems

Description: On-spacecraft computation needs are expected to grow significantly as future architectures plan to leverage machine learning (ML) and artificial intelligence (AI) algorithms for more rapid onboard data processing, command and control, and other dynamic decision-making. Management of waste heat is a significant challenge that may limit AI/ML capabilities due to the difficulties of rejecting heat in the space environment through radiative means only. While sophisticated active and passive heat transfer technologies exist to move heat away from heat-generating components, they are ultimately limited by the means of rejecting that thermal energy into space. (i.e. discrete radiators or components of the spacecraft bus) This topic seeks novel designs and approaches for technologies that can significantly improve the thermal power flux density (W/m2) radiated into the surrounding space environment compared to radiators used on today's spacecraft without significantly compromising normal spacecraft operations. These approaches should be broadly applicable to all orbits and space environments. The use of typical GPUs and/or NPUs for edge AI/ML computing applications can be assumed when determining thermal management needs. Proposed approaches should be designed to be ultimately deployed from a launch vehicle and survive in a space environment.

Example Approaches: (nano)materials engineering, device engineering, directional emissivity, multilayer films

ORISE Topic Number: ICPD-2026-06-Quiescent

Title: Practical Thermoradiative Devices for Quiescent Power

Keywords: thermoradiative cells, solar cells, inverse photovoltaics, remote power, semiconductors, semiconductor materials, photodiodes, epitaxy, III-V, MOCVD, thermal radiation, emissivity, radiative cooling, heat transfer

Description: Due to their inherently passive nature, (no noise or moving parts) thermoradiative devices could be an attractive means to provide quiescent power to extend the life of remotely deployed battery-powered devices on land, water, or in space. Furthermore, in scenarios where waste heat is available, they could offer an alternative to thermoelectric devices for power generation. While published studies suggest a theoretical (ideal) maximum power density of approximately 50 W/m2, experimental demonstrations remain significantly below power densities required practical application. Limitations include the growth of high quality, low bandgap materials that maximize radiative recombination, maximizing optical outcoupling of the radiated photons, and thermal management to facilitate productive radiative processes. Approaches are sought working towards demonstrating practical thermoradiative devices for remote application with minimum power densities of at least milliWatts per m2. Projects can explore device designs and assembly, modeling, and/or materials growth, among other supporting efforts. Approaches can be suggested that can leverage waste heat available in certain scenarios that elevate the hot side temperature. (please specify applications)

Example Approaches: Directional emissive surfaces, high emissivity materials, high thermal conductivity materials, growth of low bandgap high radiative efficiency materials, device or materials structure modeling to maximize power density, thermal heat spreading

ORISE Topic Number: ICPD-2026-07-Semiconductors

Title: Quantum Point Defect Engineering in High-purity Semiconductors

Description: Quantum sensors offer tremendous potential for high sensitivity and specificity to a signal of interest. Similarly, quantum computers offer revolutionary potential for a variety of problems that are intractable using classical computers.

Key to unlocking the practical potential of both of these technologies is the ability to create a quantum state that is isolated from undesirable disruptive stimuli and can interact with either control signals (computing) or environmental stimuli of interest (sensing). The highest-performance systems for such quantum states are isolated atoms, either in a sparse atomic vapor or atom trap. Defects in a crystal can have similar quantum properties to isolated atoms, but usually suffer shorter quantum state lifetimes. On the other hand, defects in crystals offer many advantages, including higher density and dramatically more straightforward integration into practical devices. We aim to study the possibility of precision quantum point defect engineering in high-purity semiconductors.

Our vision is to develop techniques to allow the vast combinatorial space of potential host lattices and defects to be understood and exploited, such that host/defect pairs can be designed with particular properties in mind. We focus on high-purity semiconductors to take advantage of the vast research done to understand their properties and defects for purposes of conventional electronics, and the existing well-developed growth and device fabrication techniques.

Example Approaches: The history of band structure engineering is an instructive analogy. Over more than 40 years, the ability to understand and design band gaps, band alignments, band bending, etc., in layered semiconductors improved alongside the ability to grow such structures at ever greater purity and precision, leading to sophisticated engineering of high-performance quantum devices including, e.g., quantum cascade lasers, quantum well infrared photodetectors, high-electron-mobility transistors, and resonant tunneling diodes. We aim to explore if similar advances can be made in the area of quantum point defect engineering. Advances in Density Functional Theory and Molecular Dynamics codes have demonstrated success in predicting both electronic and phononic states, novel data-driven machine learning algorithms are being exploited for materials discovery, and some efforts have combined the two to make advances in high-Tc superconductor design.

We would like to investigate if a generalizable approach to quantum point defect engineering can be developed that enables hosts and defects from the vast array of unexplored combinations to be selected for applications where atomic vapors or atoms in traps have seen significant development as quantum sensors. (i.e., qubits, atomic clocks, DC magnetic field sensing, antenna-free RF sensing, THz devices, etc.).

ORISE Topic Number: ICPD-2026-08-Robust

Title: Robust and Resilient Biomanufacturing

Keywords: Biomanufacturing, scaling, bioengineering, extremophiles

Description: Biomanufacturing is rapidly emerging as a critical domain for national security. It was designated one of six critical technology areas by the Department of War in 2025 and featured prominently in the 2024 final report of the National Security Commission on Emerging Biotechnology (NSCEB). Additional major U.S. government initiatives—such as the Tri-Service Biotechnology for a Resilient Supply Chain (T-BRSC) and the Bioindustrial Manufacturing and Design Ecosystem (BioMADE)—aim to strengthen domestic manufacturing capabilities and enhance U.S. infrastructure and competitiveness. Foundational research, including efforts from SERDP and DARPA, has significantly de-risked biomanufacturing for producing complex molecules with national security relevance.

While these initiatives continue to advance scalability and efficiency, this research topic focuses on enhancing the agility and robustness of biomanufacturing systems. Industry efforts often prioritize cost-efficiency, and Department of War programs have emphasized scale and deployability. However, there remains a critical need for systems that can reliably operate across diverse and unpredictable environments. Proposals should explore novel approaches that prioritize confidence of production over scalability—ensuring functionality regardless of the manufacturing context.

Example Approaches: Agile and Resilient Biomanufacturing
Research should explore platforms that are flexible, rapidly reconfigurable, and resilient to disruptions. This includes the design and evaluation of “switchable” systems capable of pivoting between products in response to supply chain shocks or emergent security needs.

Robust and Extensible Biomanufacturing
Proposals should investigate chassis or cell-free systems that maintain performance across a wide range of feedstocks and operational platforms. Emphasis should be placed on systems that can adapt to variable inputs while consistently producing target molecules.

Extremophile-Based Systems
This area includes the identification, characterization, or engineering of novel chassis organisms capable of thriving in extreme conditions. Examples include thermophiles with extracellular electron transfer capabilities or systems that utilize nontraditional energy sources to sustain production under high temperature, pressure, or other challenging environmental factors.

ORISE Topic Number: ICPD-2026-09-NextGen

Title: Enabling Next Generation of Wearable AI: A Cross-Layer Approach to Ultra-Low Power System-on-Chip

Keywords: Wearable AI Ultra-Low Power System-on-Chip (SoC) Cross-Layer Optimization Edge AI Microelectronics Algorithmic Compression

Description: AI devices are rapidly advancing, driven by innovations in integrated circuits (ICs), algorithms, and computer architecture. These developments enable audio and video processing in mm-scale, unobtrusive edge devices, opening the door to new applications in health, fitness, and ambient intelligence. However, realizing truly wearable AI poses significant challenges, especially with respect to power consumption and system integration: battery capacity and chip area must be drastically reduced (e.g., sub-mW operation and <5mm² die size), all while supporting efficient inference using state-of-the-art AI models.

These innovations aim to deliver a compact, extremely energy-efficient SoC for real-time audio and video inference, supporting long-duration, untethered operations within flexible mm-scale wearable devices.

Three key technical innovations must be optimized and co-designed to achieve the size and energy consumption necessary for wearable edge-AI solutions for the intelligence community.
- Memory Integration: No off-chip memory to remove a primary bottleneck for size and energy.
- Algorithmic Compression: Advanced model compression strategies, aggressive quantization, parameter decomposition, and efficient approximation techniques.
- Circuit Optimization: An architecture utilizing sub-threshold circuits and dynamic power gating.

Example Approaches: To achieve the necessary power and area reductions, approaches should employ a cross-layer co-design strategy that tightly couples algorithmic innovation with hardware implementation.

- Memory Integration: Ultra low-powered on-chip memory such as MRAM arrays stacked using monolithic 3D integration could drastically reduce the size and power requirements. To manage data flow efficiently, a hierarchical buffering scheme could be utilized where specific weights required for current neural network layer are moved from MRAM blocks to high-speed local SRAM caches.
- Algorithmic Compression: Advance AI model compression could utilize sub 4-bit quantization, extreme knowledge distillation, adaptive block low-rank parameter decompositions, and sparse low-rank activations to reduce computational complexity and model size by at least 5x.
- Circuit Optimization: To achieve ultra low power, an example approach would involve dynamically power-gated Multiply-Accumulate (MAC) arrays for precision power control. Additionally, the team could implement voltage islands, allowing circuits to operate in the
sub-threshold region for maximum efficiency, while memory blocks remain at slightly higher voltage to ensure data stability.

ORISE Topic Number: ICPD-2026-10-Nanowatt

Title: Nanowatt-Scale Autonomous Computing for Battery-Less Operations

Keywords: Ultra-Low Power Electronics Energy Harvesting Nanowatt Computing Edge Artificial Intelligence Microelectronics Sub-threshold Circuit Design Power Management Units (PMU)

Description: This research initiative addresses the critical energy constraints limiting the deployment of persistent, maintenance-free edge intelligence systems in remote or inaccessible environments. The fundamental problem is the significant disparity between the power required for sophisticated processing (programmable control, edge inference, signal analysis), and the minimal energy available through environmental harvesting. To eliminate the logistical burden of battery replacement and ensure long device lifetimes, Application-Specific Integrated Circuits (ASICs) must operate solely on harvested energy, which often fluctuates and dips into the nanowatt range. Therefore, the core objective is to develop circuit-level innovations that drastically reduce active power consumption targeting tens of nano-watts and lower, enabling complex digital and analog capabilities to function reliably within a strictly harvested energy budget.

Example Approaches: To achieve ultra-low power consumption, approaches would likely employ sub-threshold or near-threshold voltage design techniques, operating transistors below their standard switching voltages to minimize active energy usage. Architectural approaches might involve designing custom, domain-specific digital accelerators that are optimized for specific edge inference tasks, offering vastly superior efficiency compared to general-purpose processors. Furthermore, approaches may explore aggressive power management strategies, such as fine-grained power gating and the use of asynchronous logic to eliminate dynamic power waste during idle periods. Finally, the solution would likely require the integration of highly sensitive Power Management Units (PMUs) capable of running on minute energy sources and efficiently regulating voltage for wake-up receivers and sensor interfaces without incurring significant overhead.

ORISE Topic Number: ICPD-2026-11-SmallMolecule

Title: Small Molecule Sensing and Detection using Synthetic Cells

Keywords: Synthetic Cells, Chemical Detection, Synthetic biology, Artificial Intelligence, Protein modeling

Description: Recent advances in research in artificial cells and synthetic bioorthogonal membranes provide the ground work for advancing synthetic cells to a point of applications. We seek new research to advance the application of chemical sensing using synthetic cells. The research should focus on quick and accurate analysis including a digital or colorimetric sensing output for small molecule detection. Diffusion though the membrane might be an important aspect to sensing in synthetic cells but binding of the small molecule to the outer membrane surface might increase speed of analysis. The synthetic cells can act as small microreactors where a binding or recognition event occurs possible through a promoter mechanism. Readout from those binding events should not be dependent on external equipment. Synthetic biological approaches are expected with possible use of gene circuits or quorum sensing, but not needed.

What are new methodologies for small molecule detection using synthetic cells? How can we visualize a reaction in synthetic cells without the aid of equipment? What is the functional genome needed for detection and identification? How can these synthetic cells sustain themselves for sensing applications over time?

Example Approaches: As an example, synthetic cells have shown promise for applications in biosensing and small molecules might be easier because of faster diffusion properties. (Majumder 2017).

An example of synthetic biology strategies for color change in cells, Kyoungwon et. al. demonstrated the production of natural colorants that could be used to provide a visual signal for molecular recognition events. (Kyoungwon 2025)

ORISE Topic Number: ICPD-2026-12-EconomicStatecraft

Title: Economic Statecraft: Tools and Methodologies for Risk, Impact, and Orchestration

Keywords: Economic Statecraft; geopolitics and geoeconomics; risk analysis; escalation; decision-support

Description: Economic statecraft is the use of economic tools to pursue foreign policy goals in times of competition, conflict, or crisis. Economic statecraft tools include, but are not limited to, those shaping trade, investment, and monetary policy. While certain tools of economic statecraft like sanctions and tariffs have received significant attention, the use, calibration, and orchestration of the full suite of economic tools are not as well understood. Further, advancing research on why and how actors select and sequence specific economic statecraft tools in dynamic bilateral and multilateral relationships would be beneficial. Progress on the below research topics would improve the ability to assess, select, and measure optimized economic tools for national strategic objectives, as well as providing new methods to determine how foreign nations may organize and select economic statecraft options, and perceive their use, refracted through the prism of culture, history, strategy, and game theory. Research should aim to develop optimized economic statecraft option orchestration and explainable justification and effect (1st, 2nd, and 3rd order) elaboration across a diverse range of geopolitical relationships and interdependencies.

Example Approaches: A non-exhaustive list of example topics for which research could aim to develop solutions:

Techniques to elucidate the strategic logic underlying selection of a suite of economic tools of national power, and provide recommendations for improved effect-oriented policies with justifications

Methods to better optimize their orchestration of economic statecraft tools in service of producing effects and managing risk

Approaches to characterize limitations, key assumptions, and constraints of foreign
adversaries’ use of economic statecraft tools, and defense against US, ally, and partner tools

Approaches to better integrate nuanced foreign policy, history, geopolitical expertise, and open-source indicators with mathematical approaches to advance geopolitical risk measurement techniques

Methodologies or technologies to characterize economic tools of national power and offer options for their effective use

ORISE Topic Number: ICPD-2026-13-Toolbox

Title: Synthetic-Biology Toolbox for Biomanufacturing

Keywords: Synthetic Biology, Bioinformatics, Database Development, Computer Science, Molecular Biology

Description: This research opportunity focuses on the creation of a comprehensive and curated database, termed "BioFactory," specifically designed to house synthetic biology gene sequences relevant to biomanufacturing processes. Currently, accessing and utilizing proven genetic components for biomanufacturing can be fragmented and time-consuming. BioFactory will address this bottleneck by providing a centralized repository of well-characterized genes, promoters, terminators, and other regulatory elements known to enhance or enable the production of valuable biochemicals. The database will feature intuitive search functionalities, detailed annotations including performance metrics (e.g., expression levels, product yields), host organism compatibility, and literature citations. Furthermore, it will incorporate a user-friendly interface allowing researchers to easily browse, compare, and download sequences for direct implementation in their own biomanufacturing projects.

The development of BioFactory offers several potential avenues for exploration. One possible approach involves systematically surveying existing literature to identify and compile a comprehensive collection of gene sequences exhibiting proven utility in biomanufacturing applications. Another avenue focuses on meticulously curating, standardizing, and annotating these sequences using a variety of bioinformatics tools. This could encompass enriching the sequences with functional descriptions, predicted protein structures, associated metabolic pathways, and known limitations to improve user understanding. A further avenue explores designing and implementing a robust, scalable, and user-friendly database architecture to support efficient data storage, retrieval, and accessibility, while considering modularity for accommodating future data expansion and evolving user needs.

The ultimate goal of BioFactory is to accelerate innovation and standardization within the biomanufacturing field. By providing a readily accessible and well-characterized toolbox of synthetic biology gene sequences, this database will empower synthetic biologists to rapidly design, build, and test novel biomanufacturing processes. This will reduce the time and resources required for de novo design and optimization, allowing researchers to focus on more complex engineering challenges and ultimately paving the way for the development of more sustainable and efficient biomanufacturing workflows. The database will also enable the straightforward replication of proven biomanufacturing processes, fostering reproducibility and accelerating technology transfer from research labs to industry.

Example Approaches:  • Literature Mining and Expert Curation: Systematically search scientific publications and patents for gene sequences related to biomanufacturing. Experts in the field will then review and curate these sequences, ensuring accuracy and relevance.
• Standardized Metadata and Annotation: Implement a standardized schema for annotating each gene sequence with comprehensive metadata, including function, performance metrics, host organism compatibility, and associated pathways. This will enable efficient searching and comparison.
• User-Friendly Web Interface: Develop an intuitive web-based interface for users to easily search, browse, and download gene sequences. This will include advanced search filters, visualization tools, and submission forms for community contributions.
• Community Engagement and Collaboration: Foster a collaborative environment by encouraging users to contribute new sequences, annotations, and feedback. This will ensure the database remains up-to-date and reflects the evolving needs of the biomanufacturing community.

ORISE Topic Number: ICPD-2026-14-Crustal

Title: Leveling and merging crustal scale marine and airborne magnetic surveys

Keywords: marine magnetic surveys, airborne nagnetic surveys, magnetic data processing, diurnal correction, crustal-scale magnetic anomaly grids

Description: This research topic focuses on the analysis and enhancement of data leveling techniques for marine magnetic surveys. The core objective is to assess the efficacy of current leveling correction approaches and to innovate new methodologies specifically designed to mitigate leveling errors in marine collected magnetic data. The research will investigate the unique challenges presented by marine surveys, such as extended data collection periods and the difficulty of effective diurnal variation monitoring, to improve robustness of crustal magnetic anomaly maps. Crustal magnetic anomaly maps, which have shown promise for an alternative positioning and navigation (Alt PN) method, are compromised by significant data inaccuracies known as leveling errors. These errors primarily arise in marine magnetic surveys because, unlike faster airborne surveys, their prolonged duration makes it impractical to effectively monitor and correct for the diurnal variation of the Earth's magnetic field. Existing correction methods, often adapted from airborne surveying, are insufficient for the unique temporal and spatial scales of marine data collection. Consequently, persistent leveling errors degrade the quality of magnetic anomaly products, creating a need for new, robust data leveling techniques tailored to the specific challenges of marine magnetic surveys.

Example Approaches: This research will be executed through a phased methodology designed to systematically address and resolve leveling errors in marine magnetic data.
1. Assessment of Existing Techniques: Involve a comprehensive review and quantitative assessment of current data leveling techniques. Using a curated selection of historical and contemporary marine magnetic datasets, we will benchmark the performance of standard methods to precisely identify and characterize the nature and magnitude of residual leveling errors. This baseline analysis will establish the current state-of-the-art and its limitations.
2. Development and Innovation of Novel Methods: Concentrate on the development of novel correction algorithms. This effort will investigate and expand upon techniques that have already shown considerable promise, such as leveling methods based on line-to-line correlations and the application of weighted spatial averages and temporal filters. The primary goal is to design and implement new workflows specifically tailored to the unique challenges of marine magnetic surveys. As data leveling errors in marine surveys are a persistent challenge, proposals and research focused on new techniques to eliminate or minimize these errors will be prioritized.
3. Validation and Integration for Operational Support: Newly developed methods will be rigorously validated against the benchmarked datasets. Success will be quantified by the demonstrable reduction in cross-line leveling errors and the overall improvement in the internal consistency of the magnetic anomaly data. If successful, the new leveling approaches and workflows will be structured to be complementary to ongoing research efforts on processing
"raw" magnetic data.

ORISE Topic Number: ICPD-2026-15-Novel

Title: Utilization of novel materials for investigation of sensor capabilities

Keywords: HTS: high-temperature superconductors, LTS: low-temperature superconductors, Chip-scale, Wafer manufacturing of novel sensors, Quantum sensors.

Description: Superconducting materials show promise for multimodal sensors due to their quantum phenomenology and multiple mechanisms for detection. However, due to the tradeoffs between operating temperature (cryocooling) and ease of circuit fabrication (manufacturability), the choice of materials becomes critical. For example, high-temperature superconductors (HTS) are much easier to deploy than low-temperature superconductors (LTS) from SWaP perspectives, but they require different fabrication methods. The research will investigate the range of superconductors with specific interest in a-axis YBCO due to its unique combination of high operating temperature and manufacturability. The study of superconducting nanowire single photon detectors (SNSPDs) offers a good opportunity to compare material properties of families of LTS (such as niobium and its variants) and HTS (such as REBCO variants including and especially a-axis YBCO). Metrics of interest include, but are not limited to, transition temperature (T_c), critical current density (J_c), penetration depth (ë), electronic relaxation (ô_e), thermal relaxation (ô_th), kinetic inductance (L_k), dark count, and jitter. Counting photons is also relevant in addition to detecting single photons. A comparison of SNSPD geometries and their manufacturability in the different material sets will also be useful.

Example Approaches: Material choice is especially important given the potential of superconducting microelectronics to implement multimodal systems-on-chips, including computing at the edge within the same cryo packages as the single photon detectors, photon counters, magnetic detectors, radiofrequency sensors, and others.   

ORISE Topic Number: ICPD-2026-16-Entanglement

Title: Entanglement – Enhanced Rubidium Magnetometry for Quantum Metrology

Keywords: Quantum Entanglement, Quantum Magnetometer, Rydberg atom, Quantum Metrology, Quantum Sensing.

Description: This research aims to develop entanglement-enhanced techniques for rubidium-based magnetometry, with the goal of surpassing the Standard Quantum Limit (SQL) in magnetic field sensing. While conventional rubidium magnetometers are constrained by quantum noise, utilizing entangled atomic states or photon entangled states as probing mechanisms has the potential to significantly enhance sensitivity, enabling measurement precision beyond that achievable with classical methods. The primary challenge is to generate and sustain highly entangled states in rubidium ensembles within a stable experimental framework, while effectively minimizing decoherence caused by external noise.

To address this challenge, the research will focus on designing a robust and scalable platform for creating and maintaining entangled states in rubidium atoms. This includes optimizing coherence time and implementing advanced shielding techniques to protect the system from environmental disturbances. The goal is to push the limits of quantum sensing, offering more precise magnetic field measurements and contributing to the advancement of quantum metrology

Example Approaches: To address the challenge of surpassing the Standard Quantum Limit (SQL) in rubidium-based magnetometry, this research proposes a carefully constructed experimental setup. The process involves cooling and trapping rubidium atoms using laser-cooling techniques, followed by the generation of spin-squeezed states through quantum non-demolition measurements or off-resonant optical pumping. To maintain the coherence of these entangled states, advanced techniques such as magnetic shielding and optical lattices will be employed to minimize the effects of external noise and atom-atom collisions. Once the entanglement-enhanced magnetometer is constructed, its sensitivity will be systematically tested and compared to that of traditional rubidium magnetometers.

ORISE Topic Number: ICPD-2026-17-Prediction

Title: AI/ML Prediction of Atmosphere Effects on Optical Signals

Keywords: AI/ML, communications, atmosphere, space

Description: As communications in GHz frequency bands become more crowded, the telecommunication industry constantly works to improve THz communication infrastructure. The inherent increase in data rate and bandwidth that comes with THz communications increases the flow of information. This is particularly important to data centers. The current industry push to establish data centers in space requires accurate prediction of atmospheric effects to achieve the best performance.

The fact that the atmosphere is constantly changing has made it extremely difficult to predict on time scales and accuracy needed for reliable, high-data-rate communications. The objective of this program is to explore how well AI/ML techniques can predict atmospheric effects on optical signals.

Example Approaches: Example approaches may range from historical data analysis to more model-based approaches.

ORISE Topic Number: ICPD-2026-18-Stochastic

Title: Application of stochastic resonance in sensing and communication

Keywords: Stochastic resonance, signal amplification, noise, photonics, quantum, IR, communication

Description: Stochastic resonance leverages noise to amplify weak signals. This phenomenon enables various sensory processes in both biological and engineering systems. Moreover, by overcoming the limitations of thermal noise, stochastic resonance can facilitate the development of novel communication modalities.

This topic welcomes proposals in different areas of sensing and communication, in particular for applications in the quantum regime or IR sensing. Stochastic resonance may enable detection of weak signals below the thermal noise potentially opening the door for IR sensing without the need for cryo-cooling.

Example Approaches: An experimental demonstration of increased detection sensitivity in an IR system using stochastic resonance.

An experimental demonstration of low SNR signal transmission and detection in the quantum regime using stochastic resonance.

ORISE Topic Number: ICPD-2026-19-Excitation

Title: Nuclear Clock Transition Excitation Source

Keywords: Nuclear Clock, PNT, Thorium-229, VUV Laser, Timing, Quantum Sensing

Description: The discovery in 2024 of the nuclear transition energy in 229Th offers a breakthrough opportunity in Position, Navigation, and Timing (PNT) research. This isomeric transition offers unprecedented stability and precision, far outstripping state-of-the-art atomic clocks. A vital component to any optical clock system, the light source often represents the largest size, weight, and power (SWaP) draw among the clock components. In particular, nuclear clock research has traditionally leveraged optical frequency combs, distributing power over a wide bandwidth. For this project, a successful applicant will propose a research program to develop an extremely narrow linewidth, continuous wave light source capable of exciting the aforementioned nuclear transition in a low-SWaP package consistent with spaceflight.

The successful applicant will perform literature review to identify the most promising approaches to producing the robust, low-SWaP laser system. Having selected an approach, they will demonstrate the relevant mechanism(s) of action at the lab benchtop level. The ideal candidate will bring expertise in laser physics, optical engineering, and atomic spectroscopy to this interdisciplinary project.

Example Approaches: (1) Integrated Photonics
Beginning with a low-powered seed laser in the optical or infrared wavelengths, implement cascaded frequency doubling processes using lithium niobate waveguides. This approach would require advances in precision engineering to ensure highly efficient conversion at each stage, and material research to produce integrated waveguides capable of efficiently transmitting at ~148.4nm.
(2) Direct UV generation
Using quantum cascade laser technology, manufacture semiconductor heterostructures with quantum wells capable of emission at ~148.4nm. This approach would likely focus on techniques for the precision manufacturing of the quantum wells operating in the vacuum ultra-violet (VUV) spectrum.
(3) Atomic vapor
Using an atomic vapor system with appropriate energy levels and nonlinear susceptibilities, engineer a vapor cell to optimize phase-matching conditions to ~148.4nm light via multi-wave mixing. This approach would require modeling of the multi-level atomic system and an experimental demonstration of the vapor cell.

ORISE Topic Number: ICPD-2026-20-Light

Title: Structured Light

Keywords: Squeezed light, structured light, optical communication, optical wavefront, atmospheric compensation, optics

Description: Sending an optical signal through the atmosphere presents significant challenges. Atmospheric turbulence will distort and weaken the signal. Adaptive optics can ameliorate this problem somewhat, but at significant hardware cost. For this project, a successful applicant will propose a research program to explore methods of using structured light to design self-healing wavefronts – i.e., beams that disperse less than a Gaussian beam under turbulence conditions.

The successful applicant will identify a strategy for advancing the state of the art and demonstrate a mechanism of action – whether via benchtop experiments or in simulation.   

Example Approaches: (1) Simulation
Define distinct classes of non-Gaussian beams and model their propagation through the atmosphere under a variety of rational conditions. Use the results of the simulation to identify which beams are optimal under various classes of conditions.
(2) Metasurfaces
Propose a metasurface capable of tunable self-healing beam generation. Perform fabrication, identifying design trades, and demonstrate the metasurface light generation capabilities in a bench top experiment. Optimize performance and identify capabilities and limitations.
(3) Hybrid beams
Investigate structured light beams that carry both classical and quantum information through atmospheric channels by engineering wavefronts with spatially separated regions optimized for different encoding schemes. In this way, demonstrate either at the benchtop level or in silico a technique for maintain quantum coherence while simultaneously maintaining a classical data link.

ORISE Topic Number: ICPD-2026-21-Advanced

Title: Graded property RF structures using Advanced Manufacturing

Keywords: radio frequency, RF, graded index, 3D printing, advanced manufacturing

Description: Problem: Radiated RF signals are vital to defense communications. Efficiently receiving or transmitting weak signals at low frequencies using small antennas is a persistent challenge.3D printing of RF materials has made it possible to manufacture RF lenses/absorbers with gradients in properties, potentially advancing performance or reducing the size of antennas. This research should holistically study novel structures and novel manufacturing methods / materials for optimal performance.

Example Approaches: 

1. Fused filament fabrication with varying densities of the lattice of an RF polymer or conductive absorber
2. Sterolithography of varying lattice densities using Ceramic loaded photoresin
3. Actively mixed 2-component syringe extrusion of two materials

ORISE Topic Number: ICPD-2026-22-Laser

Title: Surface Chemistry of Short-pulse Laser Ablation on Exotic Material for Circuit Fabrication

Keywords: laser direct structuring, surface chemistry, laser ablation, nongalvanic plating, emerging technologies

Description: Problem: Next generation microelectronics in defense systems will include heterogeneous integration (HI) of different technologies, and face challenges of SWAP.

Using short-pulse laser catalysis, it is possible to enable HI and create 3D packaging, but more capability lies in process success on better RF and thermally dissipating substrates, and non-galvanic plating of metals other than copper. This research should advance understanding of laser parameters, substrates, and plating chemistry for a successful process.

Example Approaches: Picosecond infrared lasers on Copper Chomite doped polymers
Nanosecond green lasers on ligh-colored alumina and Sapphire, SiC, and AIN.
UV Excimer lasers on Aluminum Nitride to form continuous conductive films sans plating
Non-galvanic autocatalytic plating of other metals (Aluminum, Ni, others?)
Factors (wavelength and pulse width, material properties) contributing to minimum feature size

ORISE Topic Number: ICPD-2026-23-Optoelectronic

Title: Novel optoelectronic devices for classical computing and quantum sensing

Keywords: Light-matter interactions, strong coupling, optoelectronics, quantum sensing, neuromorphic computing, quantum emulation, two-dimensional materials, optical cavities, nanotechnology

Description: Problem Statement: The IC requires optoelectronic hardware for future computing and quantum sensing which is compact, low power, and operational at room temperature.
Topic Description: Solid-state systems with strong light-matter interactions display emergent quantum behavior and highly nonlinear responses. This topic aims to advance the use of
low-dimensional materials in strongly-coupled optoelectronic devices where electronics are used to tune the coupling and/or transduce changes in the optical cavity with external perturbations.

Example Approaches: Approaches can include fabrication, characterization, and/or modeling of low-dimensional materials coupled to optical cavities with integrated electrodes.

ORISE Topic Number: ICPD-2026-24-Neuromorphic

Title: Visual language model that will be able to leverage neuromorphic and quantum for HSI

Keywords: large language model, neuromorphic computing, Quantum computing, hyperspectral image (HSI)

Description: (U) Hyperspectral image understanding for foundation data generation is a challenging problem for earth observation mission, current method usually only work for electro-optics images, and less efficient for high dimension data. We are looking for non-traditional computing approach that out-perform traditional GPU computing, which requires innovation on tokenization scheme.

Example Approaches: Neuromorphic computing is one example approach that leverage spiking neural network and amplitude encoding. We will leverage feature learning and use traditional rate model as teacher in the training process so that we can develop spiking neural network that can be deployed to neuromorphic chip. Meanwhile, quantum computing is another approach that we can explore using quantum circuit, which can leverage quantum computer's capability.

ORISE Topic Number: ICPD-2026-25-Microelectronics

Title: Microelectronics for Energy Harvesting, Power Management, and Wireless Sensors and Communications

Keywords: Power electronics, power management, low power microelectronics, non-linear antennas,
electromagnetic field harvesting, wireless sensors, wireless communication

Description: The Intelligence Community (IC) has a pressing need for low power energy harvesting, power management, and communication electronics to meet mission requirements in large-data, low power applications. The development of new microelectronic designs is critical to further decreasing the size and efficiency of the needed architectures. To further improve the competitive advantage of the IC and develop the next generation power electronic and communication technology, three major initiatives have been proposed:
1. Multi-Input Energy Management Electronics
The development of highly efficient electronics that can accommodate multiple input sources, such as solar panels or other energy harvesting devices, is essential for optimizing energy management in modern low-power applications. At the chip level, circuit designs need to seamlessly integrate these diverse energy inputs, often with varying voltages and power levels, to charge and discharge a battery while simultaneously powering a load. For this research topic, a prototype power management controller that can continuously power a load from a combination of uW-mW harvested energy inputs with a battery acting as a reserve power source is desired.
2. Non-linear, Near-Field Antennas for Power Line Communications
The capability to receive data over power lines would enable data transmission in systems without dedicated fiber optic cables. This proposal topic aims to model, design, and prototype a near-field antenna to wirelessly receive data signals transmitted via a power line. Proposals of interest would model and design antenna and microprocessor topologies for the wireless detection of both narrowband and broadband signals traveling on power lines. Non-linear antenna designs are of strong interest to optimize low frequency communications on electrically small antenna systems under the most stringent low signal power conditions. For prototype designs, factors such as small size, useful voltage output (typically 3.3V-5V), placement orientation (clamp around vs probe sitting on a wire), and the overall design and manufacturability of the system should be considered. Project outcomes should include EM system models of power lines, system models for energy harvesting, and preferably a prototype microelectronic system for energy harvesting.
3. Electromagnetic Energy Harvesting from Motors
In modern electric motors with variable frequency drives, a common failure mechanism can arise from the electromagnetic interference (EMI) generated from high switching frequencies and voltage pulses. This EMI can manifest as currents that can leak and propagate through the motor frame, cables and shielding, or any other conductors, steel, or aluminum parts of the structure. This research proposal aims to study the topology of EMI currents in motor systems to develop and optimize a microelectronics energy harvesting process while minimizing damage from the normally undesirable EMI currents. By adding energy harvesting microelectronics, wireless engine monitoring sensors can be optimized while minimizing wiring harness requirements.

Example Approaches: 1. Proposals of interest would address advanced power management circuitry capable of efficiently converting and regulating input voltages typically ranging from sub-volt levels to 3.3V or 5V to continuously power a load from a combination of uW-mW harvested energy inputs with a battery acting as a reserve power source. Efficient operation across this range is crucial to maximize energy utilization, extend battery life, and ensure reliable powering of devices in applications such as IoT devices, portable electronics, and renewable energy systems.
2. Proposals should design, build and optimize a non-linear antenna to receive low signal power data on a power line optimized for a given set of bandwidth, signal distance, and external noise. Factors such as detectability at range, transmission sources, size, orientation, voltage output, and overall manufactuability should also be addressed.
3. Proposals should model and map motor/generator topologies to analyze the electromagnetic field in conjunction with the design of a coupled energy harvesting system. Proposals should discuss models and designs as a function of magnetic or electric field density and metrics reflecting small size, low voltage output, power delivery, and design/manufacturability are preferred. The impact of distance from emanating magnetic field source, interference from out-of-phase parallel sources, and how these changes with separation distance should also be examined.

ORISE Topic Number: ICPD-2026-26-PassiveRadar

Title: Scene Characterisation Using Passive Radar

Keywords: Passive Radar · Signals of Opportunity (SoOp) · Emerging RF Sources · RF Scene Characterisation · Urban Remote Sensing · Through-wall Sensing · RF Propagation · Infrastructure Validation · Disaster Response · Wireless Channel
Exploitation · Passive Radar · Signals of Opportunity (SoOp) · Emerging RF Sources · RF Scene Characterisation · Urban Remote Sensing · Through-wall Sensing · RF Propagation · Infrastructure Validation · Disaster Response · Wireless ChannelExploitation

Description: Characterising scenes in dense urban environments remains a challenge when conventional active sensors, such as LiDAR or SAR, are constrained by deployment practicalities, regulatory limits or the need for covert operation. Passive radar systems, which make use of existing ambient radio frequency (RF) transmissions, known as signals of opportunity (SoOp), offer a compelling alternative. These signals may include VHF broadcast radio, digital television, Wi-Fi, cellular signals (4G/5G/6G), smart city or campus IoT networks and other emissions likely to be present in an RF-dense, urban environment. This research will enable those responsible for the architecture or operation of government buildings to better understand and mitigate the risks from emerging passive radar sensing techniques.
We invite proposals to investigate the effectiveness of passive radar for extracting meaningful structural or environmental information from urban scenes. A particular focus is on understanding the relative merits and limitations of different SoOp including prevalent, emerging or future signal types.
Whilst the concept of using passive radar for general detection is well established, its utility for detailed scene characterisation, such as inferring internal structures, layouts or materials of buildings remains an open question. Challenges include understanding how different signal types interact with building materials, what kind of resolution is achievable in cluttered RF environments and how signal properties affect penetration, reflection and detection.

Examples of Scenes of Interest:
• Characterising a building’s structural composition and understanding its effects on signal propagation.
• Investigating the impact of a building’s internal layout and features.
• Assessing the accuracy and reliability of passive radar for measuring features’ physical dimensions.
• Exploring how infrastructure such as cables, pipes or HVAC systems affect signal propagation.
• Investigating the accuracy with which human occupancy could be determined or quantified.

Example Approaches: Proposals may pursue experimental, simulation-based or analytical approaches. Possible directions include:
• Experimental Studies: Deploy passive radar systems to test various SoOp, including 5G, DVB-T, Wi-Fi, IoT or emerging technologies in representative urban environments to evaluate their sensing capabilities.
• Simulation and Modelling: Use ray-tracing or EM modelling tools to predict signal propagation and interaction with common construction materials and urban layouts.
• Signal Processing Algorithms: Develop or apply techniques for structure inference and imaging from passive radar data, including machine learning, compressed sensing or tomographic reconstruction.
• Time-scale Performance Analysis: Investigate how sensing performance varies when systems operate in near real-time compared with scenarios allowing for long-term integration over hours to months, seeking to improve characterisation accuracy.
• Multi-signal Analysis: Comparative assessment of SoOp types in terms of penetration depth, spatial resolution and sensitivity to structural features.

Example Techniques
• Synthetic Aperture Passive Radar (SAPR): Obtaining high-resolution data on surface structures.
• Multi-static Passive Radar (MSPR): Employing multiple receivers around a building’s periphery to enhance detection accuracy of internal features and track changes from different angles.
• Machine Learning for Signal Processing: Identifying subtle structural changes within large data sets over time.

ORISE Topic Number: ICPD-2026-27-INSTANT

Title: Instant RF Lockdown: A Faraday-on-Demand System for Prisons and Secure Facilities

Description: Illegal communication through smuggled mobile phones remains one of the most persistent threats to prison safety and public protection.   
This topic explores the concept of an instant RF lockdown mechanism a system or process capable of cutting off all radio frequency communication in a designated area on demand, creating a temporary Faraday-like effect. The challenge is to design this capability in a scalable, safe, and legally compliant way that can be activated in response to security incidents, drone incursions, or intelligence-led operations.   

Example Approaches: • Engineering deployable or room-scale RF-shielding systems (e.g., flexible or switchable materials that can block signals when energised).
• Development of low-cost, localised Faraday enclosures using smart materials or metamaterials.
• Exploring electromagnetic “red button” architectures that allow instant local lockdown.
• Assessing RF leakage patterns and modelling containment zones to inform dynamic activation.
• Evaluating health and legal implications under Ofcom and HMPPS operational constraints.
• The use of RF metamaterials.

ORISE Topic Number: ICPD-2026-28-Origami

Title: Origami Antennas with Self Actuation

Keywords: Origami Antennas, Deployable Antennas, Self-actuation

Description: Origami is the traditional Japanese art of paper folding, where flat sheets of paper are transformed into intricate three-dimensional sculptures without using cuts, glue, or other materials. Having originated with simple forms like cranes and flowers, origami has evolved to include incredibly complex designs, geometric patterns and modular structures. This ancient practice combines artistic expression with mathematical principles, making it both a artistic craft and a subject of scientific study in many engineering fields. In electromagnetics, the idea of Origami antennas has recently emerged, where a 3-dimensional antenna is formed from a series of folds in 2-dimensional sheet, bringing various application benefits such as tuneability, reconfigurability and the ability to deploy the antenna.

One application that has been a particular focus of deployable origami-type antennas is space/satellite antennas, where large, highly directive antennas are folded down to reduce the size of the payload at launch and then deployed once they reach orbit (e.g., NASA's Starshade project). Other scenarios that would benefit from such an approach to antenna design include military communications and ad-hoc networks. The antenna requirements in these scenarios differ significantly when compared to satellite communications. For example, while space antennas prioritize high gain (>30 dBi) and can tolerate single-deployment mechanisms, terrestrial military antennas must balance moderate gain (10-20 dBi) with multi-band operation (VHF through Ka-band) and survive hundreds of deployment cycles. The antenna structures may be complicated and extremely sensitive to variation in physical form, requiring robust designs that maintain electrical performance despite mechanical tolerances and environmental stresses. Additionally, terrestrial applications often demand rapid deployment and redeployment capabilities, operation across multiple frequency bands, and the ability to withstand repeated folding cycles without degradation. They may have to function reliably in harsh conditions while maintaining portability and ease of use by non-specialized personnel. The inherent advantages of origami-based designs including predictable folding patterns, high compaction ratios and self-locking mechanisms make them superior to traditional telescoping or umbrella-type deployable antennas, particularly when complex aperture shapes or conformal designs are required.

Reconfigurable origami antennas leverage controlled folding states to dynamically alter electromagnetic properties across multiple dimensions. These structures can potentially achieve frequency reconfiguration by changing their effective electrical length through folding angles, enabling operation across octave bandwidths, pattern reconfiguration by morphing between omnidirectional and directive states through aperture reshaping, beam steering through asymmetric folding or rotating faceted surfaces without traditional phase shifters, and polarization agility by adjusting the orientation of radiating elements from linear to circular polarization.

However, a number of research challenges remain. Material selection presents a critical challenge where substrates must exhibit sufficient flexibility for repeated folding while maintaining stable dielectric properties, low loss tangents and minimal hysteresis effects. Conductive materials face the additional burden of surviving stress concentrations at fold lines without cracking or delaminating. The complex form of such antennas, which combine mechanical folding kinematics with electromagnetic behaviour, lead to interesting questions regarding how to design these antennas for specific scenarios. System designs must also consider the effect of the accompanying electronics and RF front end. Actuation methods remain a formidable challenge, as traditional approaches using motors or manual intervention are impractical for many applications.

Emerging solutions include shape-memory alloys that suffer from slow response times and high-power consumption, electroactive polymers with limited force generation, and magnetically-controlled origami requiring external field generators. The integration of these actuation mechanisms without compromising antenna performance or adding excessive weight and complexity is an interesting area of research. Potential solutions may take a hybrid approach, combining multiple actuation strategies and self-folding materials responsive to environmental stimuli.  

Example Approaches: We are interested in proposals that cover one or more of the following topics:
• Frequency tunable, pattern reconfigurable and beam steering origami antennas
• Design methods and techniques for origami antennas
• Antennas that fold to an extremely low form factor
• Advanced materials and manufacturing techniques for origami antennas (shape-morphing alloys, liquid metals, 4D printing) and Origami antennas with self-actuation.

References:
- S. I. H. Shah, et al., "Lightweight and Low-Cost Deployable Origami Antennas—A Review," in IEEE Access, vol. 9, pp. 86429-86448, 2021
- Y. Yang et al., "A Review of Multimaterial Additively Manufactured Electronics and 4-D Printing/Origami Shape-Memory Devices: Design, Fabrication, and Implementation," in Proceedings of the IEEE, vol. 112, no. 8, pp. 954-999, Aug. 2024
- Mishra, A.K. et al., R.F. (2024), Robotic Antennas Using Liquid Metal Origami. Adv. Intell. Syst., 6: 2400190.

ORISE Topic Number: ICPD-2026-29-Explosive

Title: Preparation, quantification and characterisation of trace explosive samples

Keywords: trace, quantification, explosive

Description: To understand the capability of explosive trace detection solutions, specific amounts of explosive material are pipetted on to swabs or surfaces for detection systems to interrogate. This allows the limit of detection (LOD) of a system to be determined but isn’t very realistic when it comes to finding real traces. The thumb print test takes a known amount of explosive deposited on a surface and then stamps a standardised thumb on to it. It is then imprinted onto different surfaces or swabs multiple times, to create more realistic residues, decreasing the residual amount with each print. The current challenge of this technique is whilst the print is more realistic, the quantity of explosive within the print is unknown. So, whilst both tests combined provide a qualitative measure for how well a trace detection system performs, ideally, we require a method that combines the quantitative nature of the LOD technique with the more realistic thumb print test.   

As such, although we have methods for quantifying the amount of explosive on surfaces this method is destructive (solvent extraction followed by GC/LC-MS analysis) and resource intensive. We are not able to quantify mass loading before testing a surface. We therefore require a quick, non-destructive, technique that could be used to determine mass loading on a range of different surfaces. It would also be useful to understand the surface coverage, crystal form, particle size and other characteristics.

Separately, but as part of the same topic, we would also be interested in investigating innovative techniques to enable more representative, realistic and reproducible trace contamination to be deposited onto surfaces for T&E purposes.   

We could provide some examples of surfaces of interest and range of mass loadings we would need capability for.  

Example Approaches: The focus of this project is to leverage emerging technology to develop a low cost, low burden solution to quantify the amount of explosive residue on various substrates. Approaches should include the development of a prototype system.  

ORISE Topic Number: ICPD-2026-30-AerialSystems

Title: Detection and Defeat of Uncrewed-Aerial Systems Using Novel Communicaiton or Navigation Techniques

Keywords: Drones, Counter-Drones, UAS, C-UAS, Detect, Track, Identify, Effector, Defeat

Description: The proliferation of drone technology has brought significant advancements in various sectors. However, this rapid growth has also introduced a range of new security challenges.

Drones are becoming increasingly capable of being used to conduct hostile and illegal activities. Drones have become a preferred tool for smuggling contraband into prisons, allowing criminals to bypass physical barriers, and deliver directly to prisoners. Small-UAS have been widely used by both sides of the Russia/Ukraine Conflict, to conduct reconnaissance and direct attacks. The conflict has driven rapid innovations in drone technology that has been widely shared online. In recent months drone sightings have caused significant disruptions at airports. These examples illustrate some of the increasing security challenges posed by drones.

Alongside increasing use, drone technology also continues to evolve rapidly. Increasingly, both commercial and home-made drones are operating on the cellular network, making use of increasing autonomy, or other novel communication or navigation techniques (such as fibre-optic controlled drones in Ukraine). These technologies are able to reduce the effectiveness of certain traditional Detect, Track and Identify (DTI) and defeat systems, and therefore, there is a growing need to investigate and support emerging research to both detect and defeat drone incidents.

This research topic has two key aspects:
1. Detection of drones operating using communications and navigation approaches beyond the traditional 2.4Ghz and 5.8Ghz WiFi bands. A pressing challenge in this domain are drones operating on the cellular network, but we are also interested in detecting drones operating autonomously or using other novel communications or navigation technologies.
2. Low-collateral defeat of such drones. Options for law enforcement to stop
(and ideally capture) a hostile drone, without introducing significant risk to the public or infrastructure.

Example Approaches: As above. This is a very fast-moving area, research will need to be adapted in line to reflect technology advances.   

ORISE Topic Number: ICPD-2026-31-CrimePrevention

Title: Behavioural Futures: Modelling Public Trust in Tech-Enabled Security and Crime Prevention/Fighting

Keywords: AI-enabled surveillance, public trust, behavioural modelling, digital, foresight, societal resilience, adaptive policy, ethical AI, scenario analysis, lawful access, crime detection, national security, technology,
prevention and detection of crime, investigatory powers, relevant legislation

Description: The increasing integration of AI and other technologies into national security and crime prevention/fighting (particularly in surveillance, border control, and digital identity systems) raises critical questions about public trust, legitimacy, and long-term acceptance. While technical capabilities advance rapidly, public attitudes are shaped by complex socio-political dynamics and can evolve unpredictably.

Recent research by CETaS (Alan Turing Institute) shows that public perceptions of technology-enabled crime fighting already vary: some people are comfortable with machine analysis, while others prefer human intervention. Public acceptance and perceived legitimacy are dynamic, shifting with exposure, incidents, governance safeguards, and communication. As seen in early online banking fraud, behaviour and risk perceptions can adapt over time, potentially moderating initial harms. The Intelligence Community would benefit from evidence-based forecasts of these changes to inform capability rollouts, communications, and safeguards.

This project has two main elements. First, it will assess the current landscape and baseline views: as criminal techniques become more sophisticated and lawful access to data is hindered by private and secure technologies, what level of machine or human intervention does the public find acceptable in preventing and fighting serious crime? Key questions include: Are there crimes where tech-enabled intrusion is more acceptable? What minimum standards do people expect for investigations and evidence collection? What is the preferred human-tech balance in detection and investigation?

Second, building on this foundation, the project will model how public perceptions of AI and other technology-enabled security measures (e.g., facial recognition, predictive policing, digital ID systems) may shift under future scenarios, such as political instability, economic shocks, or rapid technological adoption. It will explore how behavioural adaptation, like increased digital literacy or desensitisation to surveillance, might moderate perceived harms, reshape societal responses, and influence government action.

The central problem is the lack of robust, evidence-based foresight into how public attitudes toward security technologies evolve over time, and how this evolution impacts the effectiveness, legitimacy, and ethical deployment of such systems.  

Example Approaches: • Public responses: Polling, focus groups, peer review, research review, market research.
• Scenario Modelling: Develop plausible future scenarios using horizon scanning and expert input to explore how societal attitudes may diverge.
• Behavioural Simulation: Use modelling or system dynamics to simulate public responses to AI-enabled security interventions over time.
• Longitudinal Analysis: Incorporate historical case studies (e.g., online banking fraud adaptation) to understand behavioural shifts and resilience mechanisms.
• Mixed-Methods Research: Combine qualitative foresight techniques with quantitative modelling to capture both narrative and statistical dimensions of behavioural change.

ORISE Topic Number: ICPD-2026-32-Sustainable

Title: Engineering biology for sustainable power generation

Keywords: Power sources, batteries, energy harvesting, engineering biology, synthetic biology, microbial fuel cells, soil microbial
battery, electrogenic bacteria, bio-electricity, Internet of Things, low-power devices, sustainable energy, remote sensing

Description: Reliable, long-duration power remains a constraint for distributed sensing systems, particularly when deployed in environments in which battery replacement is impractical. Power sources such as lithium batteries, solar panels or harvesting from ambient energy each have limitations relating to endurance, maintenance or environmental dependence. There is therefore strong motivation to explore bio-derived, self-sustaining power systems in which naturally occurring or engineered microorganisms generate electricity directly from soil or organic matter.

Engineering-biology approaches, such as microbial fuel cells (MFCs) or soil microbial fuel cells (SMFCs), offer compelling alternatives. These systems harness the metabolic activity of electrogenic bacteria found within soil or an organic substrate, passing the electrons they release through an external circuit to generate electricity. Recent work [1] has field-demonstrated proof-of-concept systems capable of powering water purification.    

Significant limitations remain, however, including low power density, variable output, stability whilst operating under field conditions and the challenge of interfacing with IoT electronics or power-management systems. The goal of this research topic is to develop and test engineering-biology power systems that can power IoT devices sustainably under real-world conditions.   

Example Approaches: • Microbial/biological engineering [2]:  Identify, engineer or enrich electrogenic microorganisms (e.g. Geobacter, Clostridiaceae) for improved current output, stability in variable environments and compatibility with low-nutrient or waste-substrate operation.
• Electrode and materials engineering [3]:  Develop novel electrode materials (e.g. bio-char, activated carbon from biomass waste, carbon felt/stainless steel composites) and architectures (i.e. electrode spacing, surface area, stack configuration) to optimise electron transfer and durability of microbial fuel cells.
• Power harvesting and electronics integration:  Design or adapt low-voltage, low-current energy-harvesting circuits and power-management systems to better suit the output characteristics of MFCs, suitable for powering power IoT sensor nodes, data logging or wireless communications.
• Field-deployment and sustainability assessment:  Test MFC systems under realistic or tactical conditions (e.g. outdoor, variable soils and temperatures), evaluate long-term performance, maintenance needs, environmental impacts and lifecycle sustainability. Explore techniques such as stacking of cells and modular design to scale output and increase robustness.
• Application demonstration:  Implement a demonstrator IoT node (sensor or communication device) powered by a microbial battery, measure operational lifetime, reliability and feasibility in a resource-constrained scenario.

References:
- [1] Dziegielowski, J, “Development of a functional stack of soil microbial fuel cells to power a water treatment reactor: From the lab to field trials in North East Brazil”, Applied Energy (2020).
- [2] Jiang, Y-B, “Characterization of Electricity Generated by Soil in Microbial Fuel Cells and the Isolation of Soil Source Exoelectrogenic Bacteria”, Front Microbiol (2016)/
- [3] Mutuma, B, “Valorization of biodigestor plant waste in electrodes for supercapacitors and microbial fuel cells” Chemical Physics (2021)
- [4] Hess-Dunlop, A, “Towards Deep Learning for Predicting Microbial Fuel Cell Energy Output” Electrical Engineering and Systems Science (2024)
- [5] Dziegielowski, J, “Towards effective energy harvesting from stacks of soil microbial fuel cells”, Journal of Power Sources (2021).

ORISE Topic Number: ICPD-2026-33-Intuition

Title: Human Intuition and Gut Instincts in Artifical Intelligence

Keywords: Intuition, instincts, emotion, cognition, bias-reduction, information processing, decision-making,
intelligence, law enforcement

Description: Research demonstrates that professionals in law enforcement, defence, and intelligence use evolutionary ancient and implicit processes such as gut instinct and intuition in making decisions (often described in terms of sixth-sense, hunches)1-5. The reliability of these, especially of the more cognitively oriented intuition (essentially automated and pre-conscious information processing), tends to improve with time and experience. A significant risk of encouraging too great a reliance on instinct and intuition without the pre-requisite experience is the potential of encouraging ethically and professionally problematic biases, especially among novice and early career professionals. Over the past 20-years, the legitimacy of many implicit emotional and cognitive processes has been correctly challenged on this basis. During the same time, awareness of the risks of replicating human biases in AI systems grew as those technologies were refined, and significant improvements resulted, offering the potential for fast and unbiased decision making. However, given that AI is starting to play a role in law enforcement, defence, and intelligence, and that the line between professionally helpful gut instincts and intuitions on the one side and professionally unhelpful and unethical emotional and cognitive biases on the other, can be blurred even among human decision makers. How will AI replicate the helpful while removing the harmful? Further, if this is achieved, what role might AI play in training human professionals to better deploy their instinctive and intuitive capabilities without bias?   

Problem statement - AI presents opportunities in law enforcement, defence, and intelligence. These range from: 1) fast and economic data analysis to 2) bias reduction to 3) accelerated development of human capabilities. Re 1 and 2, it has been argued that AI systems introduce ‘data-driven, pattern-recognition methods that challenge the authority of subjective expertise by offering seemingly more objective, reproducible outputs’6. To counter this proposal, while AI might enhance the speed and scalability of information processing and reduce bias, by failing to model implicit human decision-making processes, it might reduce the reliability and effectiveness of these same processes, a possibility requiring some analysis. Re 3, the potential for AI to model expert human decision-making and play a role in optimizing this in human professionals is yet to be explored. There is potential for AI to be used to assist reducing the risk of cognitive biases when making risk-based decisions in both group/committee settings or following exposure of opinions of others. Could AI facilitate decision making following the input of information and consider different viewpoints more objectively, without incorporating biases?  

Example Approaches: Research would fall into three empirically discrete but overlapping strands, each at the interface of the applied behavioural sciences and digital engineering/technology.

Strand 1: Modelling instinctive and intuitive decision making in AI systems:
• Research to identify the role of gut instinct and intuition in law enforcement, defence, and intelligence, allied to an expert commentary as to how these might be incorporated into AI systems.
• Research to model gut instinct and intuition in AI systems allied to an expert commentary as to the degree to which emotional and cognitive heuristics are relied upon in law enforcement, defence, and intelligence, have been or might be accommodated.

Strand 2: Examining tensions between bias reduction and optimal instinctive/intuitive decision making in AI systems:
• Research in reducing biases in AI allied to an expert analysis of the degree to which decreased prejudices might also reduce the reliability of information processing and decision making in the relevant sectors.

Strand 3: Use of AI in training instinctive and intuitive decision making in novice (human) professionals:
• Research into expert decision making in law enforcement, defence, and intelligence, and identification of which of these skills and capabilities are traditionally trained via formal education and which are acquired via experience.
• Research in the existing or potential use of AI in the training of professional decision-making in law enforcement, defence, and intelligence – as well as in professions characterized by similar emotional, cognitive, and behavioural demands.
• A set of data-led recommendations as to opportunities presented by AI in the context of training and/or enhancing instinctive and intuitive decision making in relevant sectors.

References:
- Campeau H, Keesman LD. “You can't really turn it off”: The police “sixth sense” as cultural schema. Sociological Forum 2024;39(3):267-80.
- Carroll A. Good (or bad) vibrations: clinical intuition in violence risk assessment. Advances in Psychiatric Treatment 2012;18(6):447-56. https://doi.org/10.1192/apt.bp.111.010025 [published Online First: 2018/01/02]
- Frantz R. Intuition and behavioural economics: A very brief history. Handbook of Research Methods in Behavioural Economics: Edward Elgar Publishing 2023:321-31.
- Gigerenzer G. Expert Intuition Is Not Rational Choice. The American Journal of Psychology 2019;132(4):475-80. https://doi.org/10.5406/amerjpsyc.132.4.0475
- Stubbs G, Friston K. The police hunch: the Bayesian brain, active inference, and the free energy principle in action. Front Psychol 2024(1664-1078)
- Kingsley J, Eunice M. From gut feeling to algorithmic precision: redefining expertise in criminal profiling with ai. Unpublished manuscript, 2025.

ORISE Topic Number: ICPD-2026-34-Behavioural

Title: The role of Machine Learning (Artificial Intelligence) in Behavioural Detection

Keywords: Artificial intelligence, personnel security, insider risk, risk detection, risk mitigation, behavioral detection, behavioral indicators, hostile reconnaissance, hostile behaviour, stress, machine learning,
predictive algorithm

Description: Problem statement – How can we leverage (now and in the future) machine learning AI to detect and predict hostile national security threats? To what extent could machine learning AI be utilized for assisting in detection of behavioural indicators for national security threats, such as insider risk and hostile reconnaissance. Can AI offer a reliable technical capacity for accurate detection of behavioural indicators associated with these threats?

Description – Protective security –
As technical authority for personnel and physical security, NPSA develops security guidance to assist organisations designing, building, operating and managing the critical national infrastructure (CNI). Part of this includes guidance on how to detect behaviour that indicates a potential national security threat, including insiders acting within an organisation, or a hostile actor attempting to conduct an attack externally.

AI could offer the potential to provide technological assistance in behavioural detection relevant to national security threats, alongside human analysis. However, its capacity, efficacy, and limitations to do so remains unclear. To what extend could AI be relied upon as a tool for detection and risk management?   

In Scope:
• Machine learning – what is the capacity of machine learning AI systems to identify behaviour patterns or indicators in large data sets? What types of human behaviours can these systems detect without explicit programming of what indicators and patterns to look for? What are the limitations for these systems in detecting human behaviours? Could the data include analysis of video/visual information to identify visual patterns or indicators of hostile behaviour?
• Prediction algorithm in machine learning: What is the current capacity for prediction algorithms in machine learning to identify future risk? Could these systems be used to predict unrepeatable, low-frequency events? How much data and what quality of data would be required for reliable outcomes if applied to threat types such as detecting insider risk or hostile reconnaissance?

Aspects that could be investigated include:
• The capacity of AI to detect threat behaviours in organisational data (IT usage/access patters, and text content (HR records))
• The capacity of AI to reliably detect behavioural indicators from CCTV footage (e.g., signs of stress and anxiety, or behaviours that are outside of the ‘baseline’ for a specific environment’)

Example Approaches: • Literature review of current research on AI capacity for behavioural detection
• Primary research testing multiple AI systems capacity for detection of behavioural indicators in relevant datasets (text, IT systems access, or CCTV footage of hostile/threat actors).

ORISE Topic Number: ICPD-2026-35-Bystander

Title: Barriers to Bystander Behaviour in High Trust Work Environments

Keywords: Bystander behaviour, upstander behaviour, concerning behaviours, reporting concerning behaviour,
help-seeking

Description: Problem statement – Teams that rely on each other for protection against physical threats (armed forces, frontline police, emergency responders etc.) often have a high degree of inter-personal trust and loyalty. This can make it difficult to seek external help when team members are struggling. In some cases, individuals who are struggling may go on to engage in maladaptive behaviours that can pose a risk of harm to themselves, their team, their organisation, and wider national security (insiders etc.). How can organisations encourage staff to report behaviours of concern that they observe in their colleagues and obtain the support their colleagues need, without damaging essential trust and loyalty within the workplace?

Description – NPSA’s research and work with organisations has frequently highlighted the issue of under-reporting or a lack of intervention when counter-productive or unusual workplace behaviours are observed by employees. Such behaviours have often been seen to be pre-cursors to insider activity or welfare issues.

We are looking to understand how best to adapt our current guidance for high trust teams and organisations, such as the Armed forces, emergency responders, and police forces, to overcome barriers to reporting concerning behaviours in teams that have intense bonds of inter-peer trust and loyalty.

Aspects that could be investigated include:
• Barriers to reporting concerns about colleagues’ behaviour in suitable cohort occupations.
• Key demographic groups at highest risk from barriers to help-seeking within suitable cohort occupations.

Example Approaches: • Literature reviews on barriers to help-seeking behaviours on behalf of others in suitable cohort occupations with analysis of key themes.
• Surveys or interviews with key cohort groups to identify barriers to help-seeking on behalf of colleagues, and what factors would help to overcome these, with analysis of key themes.
• Literature review on barriers for an individual to seek help within suitable cohort occupations, and which demographic groups are least likely to seek-help, and analysis of key themes.

ORISE Topic Number: ICPD-2026-36-InsiderRisk

Title: The Future of Insider Risk - Evolving Threats

Keywords: insider risk, future threat, workforce management, hybrid working, technology, international politics

Description: Problem statement – What are the key factors that will drive insider threat in the next 5-10 years? How can organisations prepare to mitigate this threat to national security?

Description – Understanding and countering insider threats is an ongoing effort. Part of this effort is attempting to predict how insider threats might evolve in the near to distant future, to inform guidance that helps target harden against these changes. We wish to further explore how insider threat might evolve in the future given societal changes that have emerged in recent years, such as changes to working practices since the emergence of COVID-19, economic trends and differences in impacts across generations, and growing awareness of climate change.

There are several broader trends in society, such as changing ways of working, rising extremism, and employment culture that will impact how insider threat manifests and must be detected. Further research is required to assess how these trends impact Insider Threat both now and into the future, and how to adapt guidance to help prepare for changes in how these threats may manifest in the near future.

We are looking to explore current and future trends and themes across a variety of industries and thematics. This research will provide a refreshed knowledge base. The breadth of this research is essential to elicit themes impacting insider threat that we do not currently have an awareness of. Our intention is that this research will highlight some key areas to inform future research programmes.

Aspects that could be investigated include:
- Analysis of current trends and themes impacting Insider Threat. What is it that we should be caring about? What should our advice look like for the future workforce? Themes that could be included: working patters (working from home, hybrid working etc.), international conflict and politics (ideological motivations and identity), environmental change (wider societal changes and challenges), where are the risks where it comes to different generations?
- Analysis of future challenges for addressing Insider Threat. Will guidance be fit for purpose for companies created in 10-20 years out? If not, what are the key changes we need to make to assist customers in preparing for future insider threats?

Example Approaches: • Literature review exploring how socio-cultural trends impact on individual identity (international conflicts, politics, economic changes etc.)
• Primary research into key areas of staff concerns or disenchantment in the modern workplace in relevant cohort organisations
• Historical analysis exploring how key themes such as political ideology and international conflict have impacted insider cases, comparing key time periods in modern history and undertaking longitudinal trends analysis (for example, comparing WWI to WW2, WWII to Cold War era, or Cold War to modern era).

ORISE Topic Number: ICPD-2026-37-Multimodality

Title: Integrating multimodality and context to automatic language analysis

Keywords: Linguistics, forensic linguistics, applied linguistics, computational linguistics

Description: There are two main underlying problems with automatic approaches to language analysis: a lack of ability to account for context, and a lack of interpretability of language across different modalities (for example, audio, image, video, and text).
- Human communication is exceedingly context dependent1. As a simplified example, if I state that the table needs to be moved a listener will automatically use context clues to indicate whether I mean an item of furniture, or an excel style table. These might be physical context clues, or indicators from the co-text. When people are talking about sensitive, taboo, or illegal topics, this reliance on context increases even more. Automatic language tools are improving at utilizing co-text to help improve the accuracy of work, but they are still limited in the range of context that can be considered.
- A separate, but related challenge is that online communications are exceedingly important to the intelligence community, and increasingly multimodal. This might be a soundtrack which changes the intended meaning of a picture (for example a classic circus soundtrack over a social media video of me parking my car, indicating that I am not showcasing my excellent parking skills but encouraging ridicule), or an emoji pasted over the top of an image (for example a picture of snow, with a nose emoji, indicating that the post is about nasally inhaled drugs rather than snow). Cross-modal communication like this is now the norm in many groups and societies, and that is particularly the case when discussing taboo (or illegal) topics.
- Dover (2022) highlights how significant the internet and electronic communications are to intelligence communities. Automated approaches to language analysis can enable the quick triage and handling of significant amounts of data, however where they struggle significantly is with bringing together meaning from across different modes. This means that a significant amount of the communicative content risks being lost before it reaches an analyst.
- These changes in meaning provided by either the context or the different modalities might be instantly understandable to us as humans, but an automated approach that struggles to consider such aspects, will provide a severely limited output. The topic here is designed to seek ways to combat these two problems – to integrate a holistic understanding of language with automatic language approaches. The desire is that the outputs will therefore be grounded in applied and sociolinguistics and able to address language and communication in a more accurate and reliable way, considering how language actually functions.
- Context is complicated but could include any significant information about how the text was created, for what purpose, under what conditions, created by whom and for which audience is it intended. This could include: the genre of the text (e.g. an email, social media message), the field of discussion (i.e. the topic area), external events or news (e.g. the broader socio-political context), the intended audience/interlocutors (e.g. public space versus private message between friends), the shared goals of the interlocutors (e.g. planning a crime), broader topics in the text (e.g. intertextuality in a discourse community), the previous interactions and other messages with that interlocutor (e.g. knowing that a meeting relates to an earlier discussion), and the cultural context (e.g. the metaphors and stories used may be interpreted differently in different cultural contexts). Within this project it may not be possible to study all those contextual influences on the message, but priority will be given to techniques that could be applicable across these different types of context.    

Example Approaches: The exact approach will depend on the form of automatic language analyses that are being considered, though researchers will need to source their own data set(s) to show a proof of concept. An overarching example approach would be starting from a sociolinguistic or corpus linguistic perspective, and seeking to ensure that the understanding of how language works remains in the automated approaches. This is supported by the literature, most notably Grieve et al. (2024), who in their recent paper on the Sociolinguistic Foundations of Language Modeling conclude that “incorporating insights from sociolinguistics is crucial to the future of language modeling” (p17).

However the benefits of such integration has a much longer trail of evidence. For example, in a 2013 Native Language Identification challenge (sometimes called Other or Native Language Influence Detection), where participants seek to identify an author’s first language (when they are writing in English), Bykh et al. (2013) achieved a higher classification accuracy that other participants through using linguistically-informed features in their classifier. This included features such as parts of speech, lemma realisations and use of derivational and inflectional suffixes. Further work on Other Language Influence Detection for forensic linguistic purposes by Kredens, Perkins, and Grant (2019) highlights how vital an explanatory rich approach (such as one grounded in sociolinguistic explanations and features) is to analysis of language in evidential and investigative situations.

Focusing on the concurrent analysis of both the verbal and visual aspects of Instagram posts at the same time, Caple (2018) shows that taking a corpus-assisted multimodal discourse analysis can reduce partiality and enable triangulation. Polli and Sindoni (2024) look at the multimodality in hateful memes, and that the interplay between non-hateful text and non-hateful images can be used to produce hateful messages. They note that multimodality is conceptualized differently across the domains of computer science and sociosemiotics, however they also show that AI driven models can benefit from sociosemiotic insights and incorporating a multimodal critical discourse analysis approach.

More specific focused example approaches might include:
• Given a set of political speeches or news reports that happen over time in a changing context (e.g. during a conflict), how could an understanding of context improve topic modelling, document summarisation, an understanding of the evolution of events, or other forms of automated analysis? This could include (for example) the change salience of different places or people during the conflict, or a need to show strength to an audience in reaction to provocation.
• Given a set of social media posts with associated images (e.g. memes), how could an understanding of meaning and mood be better extracted from the multi-media content? For example, memes such as Wojak and Pepe the frog are often adapted quickly to express emotion reactions and humour at a given situation – how could that data be analysed alongside the text to give a more nuanced understanding of messages. Another example would be when images are used to convey instant emotional impact – for example in Daesh propaganda, CGI from video games was used to make it seem like the Eiffel tower had been attacked, or during the 2011 London riots images were shown of the London Eye on fire. Images like this may have more impact than just text messages.

References:
- Bykh, S., Vajjala, S., Krivanek, J. and Meuers, D. (2013). Combining Shallow and Linguistically Motivated Features in Native Language Identication. In NAACL /HLT 2013 Proceedings of the Eighth Workshop on Innovative Use of NLP for Building Educational Applications, 197–206, Atlanta, Georgia: NAACL/HLT
- Caple, H. (2018). Analysing the multimodal text. In Corpus approaches to discourse (pp. 85-109). Routledge.
- Dover, R. (2022). Hacker, Influencer, Faker, Spy. Hurst Publishers.
- Grieve, Jack, Sara Bartl, Matteo Fuoli, Jason Grafmiller, Weihang Huang, Alejandro Jawerbaum, Akira Murakami, Marcus Perlman, Dana Roemling, and Bodo Winter. (2024) "The Sociolinguistic Foundations of Language Modeling." arXiv preprint arXiv:2407.09241.
- Kredens, K., Perkins, R., & Grant, T. (2020). Developing a framework for the explanation of interlingual features for native and other language infuence detection. Language and Law/Linguagem e Direito, 6(2), 10-23.
- Polli, C., & Sindoni, M. G. (2024). Multimodal computation or interpretation?Automatic vs. critical understanding of text-image relations in racist memes in English. Discourse, Context & Media, 57, 100755.

ORISE Topic Number: ICPD-2026-38-Deepfake

Title: Deepfake Acoustic Profiles

Keywords: Voice Forensics, Deepfakes

Description: Some recent research ((Williams, et al., 2025) and (Schäfer, 2025)) has investigated looking at the acoustic/prosodic features of deepfakes and real audio and has identified several features that could be used to identify deepfakes. However, the data is used is not realistic/representative of potential casework. For example, utterances are short (less than ten seconds) and can be compared to clean controlled real samples etc.

Research should be undertaken to test the robustness of such methods under realistic and challenging conditions (longer samples, state-of-the-art generators, blind comparison) and to identify characteristics that makes real speakers look like deepfakes.  

Example Approaches: The research results clearly indicate that there are several acoustic/prosodic features that can be used to identify deepfakes. For example, previous research has found the fundamental frequency (F0) to be a potential indicator of deepfakes. The researchers could start with this or many other acoustic features and assess their suitability to identify deepfakes.
An approach would be to implement in automated models to identify deepfakes and compare with traditional black box detectors and determine how this approach performs under more challenging and realistic deepfake scenarios.

The desired outcome is for the researchers to identify multiple acoustic features that can be used to identify deepfakes in a variety of realistic and challenging scenarios, noting differences based on recording environment, speaker characteristics etc.

References:
- Schäfer, K. (2025). AI Got Your Tongue? Analysing the Sounds of Audio Deepfake Generation Methods. International Conference on Multimedia Retrieval (pp. 2023-2027). Chicago: Association for Computing Machinery.
- Warren, K., Olszewski, D., Layton, S., Butler, K., Gates, C., & Traynor, P. (2025). Pitch Imperfect: Detecting Audio Deepfakes Through Acoustic Prosodic Analysis. arxiv.
- Williams, E. L., Jones, K. O., Robinson, C., Chandler-Crnigoj, S., Burrell, H., & McColl, S. (2025). How Frequency and Harmonic Profiling of a 'Voice' Can Inform Authentication of Deepfake Audio: An Efficiency Investigation. JAET.

ORISE Topic Number: ICPD-2026-39-Memory

Title: Enhancing Interviewer Memory

Keywords: Investigative Interviewing, Memory

Description: Decades of research has illustrated the fallible nature of human memory, exploring implications ranging from day-to-day contexts (e.g. Schachter, 1999) through to unreliable and suggestible eyewitness testimony (Loftus, 1996; Wright and Loftus, 2008; Zaragoza, Belli and Payment, 2013).

However, research exploring the accuracy of memory recall of interviewers in the field is limited. The aims of this research are to comprehensively examine tools and techniques from but not limited to, cognitive psychology and investigative interviewing research, that could support increasing the accuracy of memory recall of interviewers in the field. A range of techniques should be explored, from techniques the interviewer can do/use following an interview to techniques that could be applied as a team approach. Additionally, this research should take into consideration the factors that can impact upon interviewer memory. The research should also consider the practicality of applying techniques and tools within a real-world context and propose research methods that would provide ecological validity to any findings. For example, in some real-world applications there may not be the option to have technology in the interview.   

Example Approaches: Research methods such as field studies, analysis of real-world data and laboratory studies have been used previously to explore how memory functions during interviewing, including what factors impact on accuracy of recall. Similar approaches could be used for this study with the focus on read-world use cases relevant to the IC and FVEY IC Partners.  

ORISE Topic Number: ICPD-2026-40-Microhaplotypes

Title: Advancing Forensic DNA Analysis Using Microhaplotypes: Enhancing Mixture Deconvolutions and Ancestry Inference

Keywords: Forensics, DNA, Sequencing, Microhaplotypes, Mixture deconvolution,
Ancestry inference

Description: Traditional forensic DNA markers such as short tandem repeats (STRs) and single nucleotide polymorphisms (SNPs) have limitations when applied to degraded samples, complex mixtures, and ancestry inference. While STRs offer high discrimination power, they are prone to stutter and mutation, and SNPs, though stable, often lack sufficient allelic diversity. There is a growing need for alternative genetic markers that can overcome these challenges and enhance forensic capabilities.

Microhaplotypes are short DNA segments composed of multiple closely linked SNPs that are inherited together as a unit, forming distinct haplotypes. These markers can be reliably distinguished using modern sequencing technologies and offer several advantages over STRs and SNPs, including low mutation rates, high allelic diversity, and suitability for degraded DNA. Their combined inheritance makes them particularly promising for forensic applications such as human identification, mixture deconvolution, and prediction of biogeographical ancestry.

Initial studies have demonstrated the potential of microhaplotype-based models to infer the biogeographical ancestry of contributors in two person mixed DNA samples. However, for microhaplotypes to be adopted in routine forensic practice, several challenges must be addressed, these include:
• The availability of sufficiently large and diverse population databases.
• The design of a robust forensic microhaplotype panel.
• The development of advanced bioinformatic tools for haplotype phasing, interpretation, and mixture analysis.

Additionally, machine learning and statistical modelling are needed to enhance predictive accuracy and automate complex analyses.   

Example Approaches: Microhaplotypes are increasingly recognised as a promising next-generation tool in forensic genetics, offering enhanced resolution for individual identification, mixture deconvolution, and ancestry inference. Their usefulness could be significantly improved through the integration of advanced bioinformatics, machine learning, and statistical modelling. Machine learning enables the detection of complex patterns in genetic data, supporting accurate interpretation of mixtures and ancestry classification. Bioinformatics tools are essential for processing sequencing data, phasing haplotypes, and ensuring reliable profile generation, including from degraded or low-template samples. Statistical modelling provides the quantitative framework for population genetics analysis, likelihood estimation, and method validation, ensuring robust and defensible forensic conclusions. Together, these disciplines will enhance the precision, scalability, and evidentiary value of microhaplotype-based forensic DNA analysis.

Examples of potential approaches:
• Identify microhaplotype panels for forensic relevance across diverse populations.
• Develop and validate protocols for microhaplotype-based mixture deconvolution using massively parallel sequencing.
• Assess the performance of microhaplotypes in ancestry inference compared to existing STR and SNP-based methods.
• Create tools and databases to support forensic laboratories in adopting microhaplotype analysis.

ORISE Topic Number: ICPD-2026-41-Ocean

Title: Ocean acoustic modelling for superior environment intelligence

Keywords: Sonar, underwater acoustics, ocean, measurements, modelling, data, information, intelligence, surveillance, reconnaissance

Description: Next Generation and Generation After Next acoustic superiority (including sonar, acoustic communications and any other acoustic technology/operation) in the underwater battlespace will depend significantly on our understanding and exploitation of the ocean acoustic environment (both the propagation of sound and the inherent noise created by the environment). Understanding acoustic behaviour in the ocean environment, and acoustically-relevant properties of the environment, is highly complex depending on a wide range of factors that change both spatially and temporally across multiple scales – even on the calmest days, the ocean is constantly changing, under the influence of a wide range of complex and dynamic ocean-acoustic factors. Presently, only the simplest acoustically-relevant properties of the environment are well described by measurements and only the simplest ocean-acoustic factors are considered in modelling.

This research topic aims to combine several research challenges to produce a plenary ocean-acoustic model that can digest complex ocean-acoustic data and generate superior intelligence about the ocean-acoustic environment, reflecting a greater understanding of acoustically-relevant properties of the environment, which can be exploited for the purposes of intelligence, surveillance, and reconnaissance, as well as commercial monitoring of the ocean. Research challenges include:
• mathematical descriptions of acoustically-relevant ocean properties, such as internal waves, eddies, and spice, covering multiple spatial and temporal scales;
• investigation of methods to provide the best self-consistent picture of the ocean acoustic environment via direct and indirect measurements (e.g. acoustic measurements);
• development of physics based, data driven, or hybrid acoustic models, including noise and propagation models, to describe acoustic behaviour in the presence of different acoustically-relevant ocean properties;
• sensitivity and uncertainty analysis and quantification, based on the quantity and quality of input environment data to ocean-acoustic models;
• investigate the computational efficiency and accuracy of different models, including different model configurations;
• development of schemes to generate, visualise, and exploit (understanding the required fidelity of) the best available description of the ocean acoustic environment;
• investigate ocean-acoustic models, and other methods, to monitor the health of, and changes to, the ocean environment.

Example Approaches: The research challenges can be approached using a mix of applied mathematics, programming, statistics, data analysis, and machine learning. Example approaches include:
• develop an underwater acoustics foundation model to understand and process ocean-acoustic data and to generate ocean-acoustic information for different applications; this could include the design and conduct of large scale data collection and preparation activities and other data collection to enable fine tuning for specific applications;
• develop new analytical and numerical models to understand and predict acoustic behaviour in a variety of different environment conditions; this could include the development of methods to synthesise a variety of acoustically-relevant properties of the environment and to represent these properties in the acoustic models;
• develop an efficient framework or architecture for combining different ocean models and acoustic models; this could include the design and development of intelligent hybrid models that optimise the combinations of models based, for example, on uncertainty or computational efficiency.

ORISE Topic Number: ICPD-2026-42-Synthetic

Title: Improving Synthetic Aperture Radar Image formation through inverse modelling and Bayesian inference

Keywords: Synthetic Aperture Radar, Volumetric SAR, Position Navigation and
Timing, Inverse Problem, Bayesian inference, SAR Interference, Urban Sensing

Description: Intelligence, Surveillance, and Reconnaissance (ISR) systems rely on high-quality Synthetic Aperture Radar (SAR) imaging to deliver situational awareness. However, standard SAR imaging methods depend on simplifying assumptions that are often invalid in complex operational environments.

This research goal is to enhance SAR image quality and tactical interpretability by formulating SAR image formation as an inverse problem with a bespoke, selectable forward model. The research will focus on developing and demonstrating novel algorithms that improve performance under realistic and challenging conditions, such as:
• Interference from additional radio frequency sources, including distributed low-power noise.
• Artifacts caused by urban structures, complex wave scattering, or layover.
• Dynamic scenes, involving moving targets or structural vibrations.
• Time and positional errors during data collection.

Applicants are encouraged to address one or more of these problem areas through proof-of-concept algorithm design and demonstration. Research should incorporate Bayesian inference methods to design and evaluate inverse solvers, considering both linear and non-linear aspects and exploring dimensionality reduction or computational feasibility techniques.

Example Approaches: Image quality should be evaluated using standard SAR performance metrics (e.g., resolution, noise-equivalent sigma-zero, sidelobe ratio) alongside subjective interpretability.   

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9160976   
In this article, the solutions for several inverse problems encountered in SAR imaging are considered, using the ARFL (Air Force Research Lab) Gotcha dataset, among others, to exemplify.   

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=10548437   
This example describes a nonlinear SAR modelling capability, which is used alongside Umbra SAR data using manual trial and error to arrive at a likely structure observed within a given scene.

https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9497317   
In this paper, several regularization methods are explored aiming to mitigate against the effects of interference in the SAR imagery. The problem is posed as a signal separation problem, and real data is used to demonstrate the results.   

ORISE Topic Number: ICPD-2026-43-Invisible

Title: Tracing the Invisible: Novel Magnetic Resonance Isotope Analysis for Forensic Footprints

Keywords: Forensics, Trace Analysis, Fingerprints, Route Attribution, Origin, Material properties, Isotopic ratios,
Nuclear Magnetic Resonance (NMR) and Imaging (MRI); Bio/Chemical Agents

Description: The use of isotopic composition as a forensic and analytical tool has become a well-established discipline across environmental science, food authentication, counterfeiting, and pharmaceutical quality control.[1] Traditional bulk isotope ratio techniques (e.g. IRMS) provide valuable insights into origin and transformation processes, but they lack the resolution to distinguish isotopic variation at specific atomic sites within a molecule.[2] By harnessing the power of Nuclear Magnetic Resonance- NMR’s atomic level precision, Position-Specific Isotopic Analysis (PSIA) offers a transformative approach, enabling the direct measurement of isotope ratios at individual molecular positions, thereby unlocking a new dimension of chemical origin and process attribution.   

Recent studies have demonstrated that PSIA can reveal subtle isotopic fingerprints linked to synthetic pathways, metabolic transformations, and environmental conditions.[3,4, 5, 6] However, for complex attribution studies such as tracing the movement of a compound through multiple production stages (e.g. Chem/Bio agent synthesis) or environmental compartments the molecular isotopic landscape (e.g. blood samples, DNAs) becomes increasingly intricate. The proposed research of position-specificity will provide unprecedented insight into origin, transformation, or contamination of samples interests to wide of intelligence services including police, national security and defence.

More specifically, this research will advance the understanding of key isotopic attributes (e.g. site-specific 13C, 2H, 15N ratios) that offer maximum insight into molecular provenance and transformation history. Furthermore, it will assess the impact of localized anthropogenic influences such as industrial solvents, synthetic reagents, or environmental pollutants that may alter or obscure isotopic signatures, either enhancing or confounding interpretive value.

Example Approaches: This research builds on a rich foundation of laboratory-based and field-informed approaches that have proven effective across disciplines such as geo-forensics, pharmaceutical sciences, food authentication, counterfeit detection, and environmental provenance research. By combining multi-analytical strategies with emerging technologies, the research goal will be to extend these methods into the molecular realm using position-specific isotopic analysis by NMR (PSIA-NMR).

PSIA-NMR offers a uniquely powerful tool for probing isotopic variation at the atomic level. When integrated with existing forensic, environmental, and biochemical frameworks, it can significantly enhance our ability to trace origins, verify authenticity, and understand transformation pathways. Unlike bulk isotope techniques, PSIA-NMR reveals subtle, site-specific signatures that often hold the key to unravelling complex histories.

The molecular targets in this research may include novel pharmaceuticals, synthetic intermediates, environmental metabolites, and even bio/chemical warfare agents. Each of these compounds carries a distinct isotopic fingerprint, one that reflects its source, method of synthesis, and environmental exposure. By analysing these fingerprints at specific atomic positions, the research will uncover layers of information that are otherwise inaccessible. The insights gained will be broadly applicable across sectors, from drug regulation and environmental monitoring to food safety and biomedical research, ultimately helping us better understand where things come from, how they change, and what stories they carry.

References:
- [1] P.H. Abelson, T.C. Hoering (1961), Proc Natl Acad Sci USA 47:623–632;
- [2] A.H. Lynch et al., (2011), Rapid Commun Mass Spectrom 25:2981–2988;
- [3] S. Guyader et al., (2019), Flavour Fragrance J 34:133–144;
- [4] D.W. Hoffman, C. Rasmussen (2019), Anal Chem 91:15661–15669;
- [5] T. Jezequel et al., (2017), Magn Reson Chem 55:77–90;
- [6] E. Tenailleau et al., (2004), Anal Chem 76:3818–3825.

ORISE Topic Number: ICPD-2026-44-Deanonymising

Title: Multi-Vector Approaches to Deanonymising Privacy Coins

Keywords: Privacy coins; Digital Data Threat; Investigation; Analysis.  

Description: The threat posed by serious organised crime, rival states, and state-sponsored proxies is amplified by the digital data threats associated with privacy-focused cryptocurrencies, a form of decentralised finance. Open-source assessments suggest that some high-end law-enforcement and intelligence agencies may have developed methods to de-anonymise traceable ledgers, which sit on more open ledgers. The techniques for de-anonymising privacy coins are far more challenging. Open-source reviews of anti-money-laundering cases and crypto suspicious activity reports indicate significant blind spots in current intelligence practices.

Training from private vendors often emphasises heuristics designed for transparent chains (e.g., Bitcoin, Litecoin, Ethereum) and tactics like coin-dusting—approaches that adversaries can easily defeat and that do not transfer well to privacy-by-design systems. Where de-anonymisation of privacy coins by private entities is reported in the open source data it is said to have relied upon combining multiple investigative vectors—such as through investigation on the dark-web, examining the timing and sequencing of transactions, on examining network metadata and other data such as “know your customer” information rather than on any single analytic technique, as might be more common with transparent chains.

In terms of dedicated consultancy activity, CipherTrace (2021) claimed some ability to deanonymise a privacy coin called Monero, but researchers have remained sceptical. Shi et al 2024, have suggested a possible vulnerability with privacy coins that shows possible routes to revealing the identity of beneficial owners. Chainanalysis and Elliptic have introduced some private sector support to investigators mixing user analytics with platform telemetry, but these remain limited.
Unanswered questions in this area include:
• What technical, procedural, and organisational vectors have been used in successful deanonymisations of privacy coins to date?
• Which combinations of vectors maximise probability of attribution under real-world constraints (legal, resource, time)?
• How do ecosystem features create choke points that investigators can exploit?
• What are the governance, ethical, and oversight implications of scaling such methods?

Example Approaches:  Design and empirically ground an optimised, management-oriented framework for targeting privacy-oriented coins and tokens that improves the IC’s investigations in these digital data threats.

This could be achieved by:
• Cataloguing and analysing all known cases of successful or partial deanonymisation of privacy coins and tokens.
• Identifying and testing high-yield combinations of exploit vectors (technical, procedural, organisational).
• Producing an operational playbook that is – through focus groups, interviews and crowdsourcing - co-designed with practitioners and which foregrounds the opportunity and risks of these approaches.
• Mapping the privacy-crypto ecosystem—the features of these protocols, user and chain custodian attributes, user dynamics, linkages to dark-web users, forums, platforms—and to provide quality visual analytics.

ORISE Topic Number: ICPD-2026-45-Cybercrime

Title: Bayesian Calibration and Infererence for Agent-Based Models of Cybercrime Ecosystems

Keywords: Agent-Based Modelling (ABM), Bayesian inference, simulation-based calibration, cybercrime, complex systems, likelihood-free inference, neural density estimation, uncertainty quantification.

Description: Cybercrime represents a complex adaptive ecosystem composed of heterogeneous actors -offenders, defenders, intermediaries, and infrastructures - interacting across digital, economic, and social domains. These interactions generate emergent systemic behaviours such as ransomware market evolution, botnet coordination, and adaptive defender responses.

Agent-Based Modelling (ABM) offers a powerful framework for representing these interactions by simulating thousands of adaptive agents, which then generate emergent criminal dynamics. However, despite ABMs’ potential for strategic foresight, their adoption in criminology and cybersecurity remains limited by a key mathematical challenge: how to rigorously calibrate and infer model parameters from sparse and noisy empirical data.

Recent breakthroughs in Bayesian simulation-based inference (SBI) - particularly black-box and neural approximate Bayesian computation developed by Farmer et al. (2024) -provide a new path forward. These methods use neural posterior estimation and density ratio estimation to infer posterior distributions directly from simulated data, enabling fast and accurate calibration of models even when likelihoods are intractable.

Each agent subsystem operates on different temporal and spatial scales, requiring scalable architectures capable of running millions of interdependent agents with asynchronous events and stochastic decision rules. Designing and implementing such a simulation demands effective application of advanced computational and data engineering methods. The challenge lies not only in building a computationally tractable model, but also in ensuring that it remains interpretable, reproducible, and adaptable as new intelligence and threat data become available.

Applying these methods to cybercrime ABMs represents a novel mathematical and computational frontier. It will require developing Bayesian frameworks capable of handling multi-level uncertainty (behavioural, structural, and observational) and defining summary statistics that capture emergent cyberecosystem properties (e.g., attack frequency distributions, market turnover, trust network resilience).

Problem Statement - How can advanced Bayesian inference methods be applied to calibrate agent-based models of cybercrime ecosystems, allowing quantitative validation, uncertainty quantification, and improved predictive capacity for complex, adaptive cyber-criminal systems?  

Example Approaches: This project may include the following methodological innovations:
• Bayesian neural posterior estimation for cybercrime ABMs: Implement neural density estimation to infer posterior distributions of behavioural and structural parameters from synthetic and real-world cyber incident data.
• Likelihood-free model calibration: Extend Doyne Farmer’s “black-box Bayesian inference” approach to criminological ABMs with heterogeneous, boundedly rational agents and discrete event dynamics.
• Cybercrime Ecosystem Modelling: Extend the agent-based model completed by Huang et. Al. to fully encapsulate the cybercrime economy and apply codified behaviours and dynamics to the agents.
• Hierarchical Bayesian frameworks: Integrate multi-level data sources (case intelligence, dark web markets, attack telemetry) within a unified probabilistic model to reconcile micro-level agent behaviour with macro-level emergent outcomes.
• Uncertainty and sensitivity analysis: Quantify how uncertainty in behavioural rules or data propagates to emergent system dynamics and predictive performance.
• Validation through synthetic data experiments: Compare model forecasts with historical cybercrime market evolutions (e.g., ransomware economies) to assess out-of-sample predictive validity.

References:
- Agent-Based Modeling in Criminology, Annu. Rev. Criminol. 2025. 8:75–95, Birks, D; Groff, E; Malleson, N.
- Agent-based modeling in economics and finance: Past, present, and future, Journal of Economic Literature, 2025, Axtell, RL; Farmer, D.
- Forecasting Macroeconomic Dynamics using a Calibrated Data-Driven Agent-based Model, M Pangallo, F Lafond, JD Farmer, S Wiese – 2024
- A survey of agent-based modeling for cybersecurity. Human Factors in Cybersecurity. 2024. Vestad A, Yang B.
- Systematically understanding the cyber attack business: A survey. 2018. ACM Computing Surveys (CSUR), 51(4), pp.1-36, Huang, K., Siegel, M. and Madnick, S.

ORISE Topic Number: ICPD-2026-46-SpeakerAudio

Title: Machine learning-based restoration of degraded speaker audio

Keywords: Speech enhancement, audio source separation, speech synthesis, intelligence, surveillance, data science, machine learning

Description: Many teams working across intelligence and law enforcement encounter ‘unclean’ audio in which the speech of interest is very difficult to discern and transcribe, resulting in sparser intelligence, frustrated operations and under-exploited data.   

While current machine learning-based methods for removing noise from unclean recordings of speech [1][2] perform well in scenarios where the sound of a single speaker is masked to some extent by environmental noise, they often fail in cases where the speaker of interest is faint and/or distorted. Such tasks become additionally complicated when multiple speakers are talking simultaneously, requiring additional processing [3] that risks further degrading the speech of interest. Existing open-source, synthesis-based approaches to the restoration of degraded speaker audio [4][5] show promise but do not yet restore the speech to the extent that it is reliably intelligible. These shortcomings exist in spite of an abundance of open-source clean and correlated noisy speech data [6].

A research project on this topic would aim to:
• Thoroughly assess the current technological landscape for speech enhancement and restoration.
• Determine and implement necessary improvements and adaptations to the above described approaches such that they more reliably result in coherent speaker audio. This stage would likely include re-training and fine-tuning existing models on bespoke data, and making modifications to existing machine learning architectures.
• Design, engineer and test new components to replace existing underperforming ones. These would target tasks like audio source separation, noise removal and speech synthesis.
• Consolidate the resulting models, architectures and components into a suite of tools for speech audio restoration.

Example Approaches: • Undertaking a comprehensive review of the current state of the art approaches to speech enhancement and restoration, including real-time approaches.
• Testing any approaches not already known to the advising agency on real-world data.
• Devising novel strategies for simulating degradations of audio
• Creating new datasets that more accurately reflect the speakers and degradations commonly found in the targeted scenarios (for example, artificially degrading an existing clean speech corpus [6] and combining it with bespoke noise data).
• Retraining and fine-tuning models in [1], [3], [4], [5] and other approaches found during the review, on speech and noise audio more closely correlated to the target use-cases.
• Design and implementation of novel approaches to speech audio enhancement and restoration; devising novel strategies for quality testing.
• Building new pipelines combining these new datasets, models and components.

References:
- [1] Defossez, A. et al, ‘Real Time Speech Enhancement in the Waveform Domain’, Interspeech 2020 paper and Python library. https://doi.org/10.48550/arXiv.2006.12847.
- [2] Sainsburg, T. et al, ‘Noisereduce: Domain General Noise Reduction for Time Series Signals’, Python library. https://doi.org/10.48550/arXiv.2412.17851.
- [3] Nachmani, E. et al, ‘Voice Separation with an Unknown Number of Multiple Speakers’, ICML 2020 paper and Python library. https://doi.org/10.48550/arXiv.2003.01531.
- [4] Liu, H. et al, ‘VoiceFixer: Toward General Speech Restoration with Neural Vocoder’, Python library. https://doi.org/10.48550/arXiv.2109.13731.
- [5] Kirdey, S., ‘VoiceRestore: Flow-Matching Transformers for Speech Recording Quality Restoration’, Python library. https://doi.org/10.48550/arXiv.2501.00794.
- [6] Dubey, H. et al, ‘ICASSP 2023 Deep Noise Suppression Challenge’, dataset. https://github.com/microsoft/DNS-Challenge.

ORISE Topic Number: ICPD-2026-47-BioInspired

Title: Bio-Inspired Molecular Sensors for Adaptive Computing and Environmental Intelligence

Keywords: Molecular Sensing · Synthetic Biology · Biomolecular Computing · DNA/RNA Logic · Chemical Reaction Networks · Adaptive Systems · Biosensor Interfaces · Low-Power Computing

Description: Biological systems have evolved remarkable capabilities for sensing, processing and responding to complex environmental signals. From molecular recognition by nucleic acids to the self-regulating feedback of metabolic networks, biology offers rich examples of computation without silicon. Advances in molecular biology, synthetic chemistry and bioengineering now make it feasible to design molecular sensors that also compute; systems able to detect signals, process information and trigger responses autonomously.

We invite proposals to investigate bio-inspired molecular sensors, integrating sensing and computation at the molecular or cellular scale. The goal is to explore how these systems can enable adaptive, low-power intelligence in settings where traditional electronics cannot easily function, such as inside living systems, in extreme environments or in highly miniaturised platforms.

Of particular interest are systems that:
• Exploit DNA, RNA or protein logic circuits to perform computation on environmental or biological inputs.
• Combine molecular sensing and decision-making, enabling autonomous activation or suppression of downstream responses.
• Operate in challenging or resource-limited environments, where conventional sensors are unsuitable.
• Provide new routes to biological-electronic interfaces, allowing molecular information to be captured and acted upon in real time.

Example Approaches: Proposals may take experimental, simulation-based or hybrid approaches. Possible directions include:
• Molecular Logic and Computation: Design and characterise molecular circuits using DNA strand displacement, riboswitches, enzymatic networks or similar mechanisms, which perform logic or classification tasks in response to chemical stimuli.
• Chemical Reaction Network Modelling: Use CRN frameworks to model molecular computation and assess the robustness, scalability and energy efficiency of different network topologies.
• Biohybrid Sensor Interfaces: Integrate molecular sensors with electronic, optical or microfluidic platforms to achieve reliable signal transduction and data readout.
• Synthetic Biology for Sensing: Engineer living or cell-free systems to detect multiple analytes and perform programmable responses. Explore how gene circuits or metabolic pathways can implement logic gates, memory or feedback control.
• Assurance and Verification: Develop methods to verify, calibrate and stabilise molecular sensing systems, ensuring repeatability, robustness and security in operational environments.
• Synthetic Cells and Partitions: Investigating physical partitions in a functional network that require transport control mechanisms that manage/coordinate system behaviour.

ORISE Topic Number: ICPD-2026-48-Transnational

Title: Mapping the Transnational Child Sexual Offending Ecosystem

Keywords: National Crime Agency (NCA)
Child Sexual Abuse and Exploitation (CSAE)
Transnational Child Sex Offender (TCSO)
Online Child Sexual Exploitation (OCSE)
Contact Child Sexual Abuse (CCSA)
Child Sexual Abuse Material (CSAM)
Sexual Exploitation of Children Travel and Tourism (SECTT)

Description: Please be aware that this is a sensitive topic involving discussions related to child sexual abuse, which may be distressing for some individuals; therefore, careful consideration should be given before engaging with this topic. We will ensure that appropriate support is identified and made available before proceeding with any related work.

A global study on sexual exploitation of children in the context of travel and tourism (SECTT) concluded that increased global tourism and greater information / mobile technology access have escalated risk in every country (expat.org.uk). The same report emphasises that transnational child sex offenders exploit situational vulnerabilities such as weak child protection in destination countries, inadequate detection, and mobility of offenders. Another report notes that the scale of transnational child sexual abuse is poorly known, but conservative estimates place the number of under 18 victims at 1–2 million per year in transnational offending contexts (Crimes without borders, 2024). The number of undetected / unreported offending is likely very large, given limits of international information sharing, victim under-reporting, corruption in destination countries, and a lack of standardised data collection. These figures support the assertion that transnational child sexual offending is a major global risk.

Transnational child sex offenders (TCSOs) present a profound and persistent threat to the safety and well-being of children worldwide. These offenders exploit the increasing ease of global travel, international connectively and the disparities in child protection frameworks across jurisdictions. Their offending often crosses borders, not just physically but digitally and legally, making detection and prosecution highly challenging. As a result, this area of offending remains significantly under-researched and poorly understood compared to domestic forms of child sexual abuse.

This research proposal seeks to address critical gaps in the academic and policy understanding of transnational child sex offenders (TCSOs), through mapping and understanding of offender typologies, contextual risk factors, and systematic weaknesses, with the aim of supporting a more coherent global response. Ultimately this will inform global efforts to identify, disrupt, and prevent the sexual exploitation and abuse of children by transnational child sex offenders (TCSOs), ensuring the vulnerable receive stronger protection and offenders face consistent accountability across borders.

The opportunities for exploration and research are fluid, to promote the incorporation of researcher skills and experience. We are open to innovative approaches and ideas which may support the national and wider foreign law enforcement community in further understanding of this threat area.   

Example Approaches: • Review differences in national legislation to identify best practices and gaps in international cooperation and coordination.
• Combining international multi-disciplinary understanding to generate a holistic understanding of the threat.
• Development of a risk assessment that considers areas such as socio-economics, cultural, and governance to identify high risk regions or travel patterns of offenders.
• Review of how international organisations, law enforcement, NGOs, and the private sector collaborate, or fail to collaborate in addressing the TCSO risk.

ORISE Topic Number: ICPD-2026-49-Security

Title: Security evaluation of system-on-chip field programmable gate array desings against remote power side channel attacks

Keywords: FPGA based, System-on-Chip (SoC)

Description: With current advancement in embedded systems, field programmable gate array (FPGAs) and CPUs are usually integrated as discrete components, and are thus implemented as separate chips on board. The processor and the FPGA communicate through an off-chip bus and may share main memory (DRAM). However, due to small footprint and lower power consumption requirements, the hardware vendors such as Xilinx and Intel have introduced heterogeneous system-on-chip (SoC) FPGA designs, which integrate both processing cores and FPGA fabric in one silicon die [1,2,4]. These integrated FPGA designs introduce new security vulnerabilities that can be exploited to perform power side-channel analysis, without physically accessing or being in the proximity of the target device [1,4].

Consider a SoC-FPGA architecture that contains both processing core (CPUs and GPUs) and FPGA fabric in one silicon die, and the system has proper protection mechanism to prevent direct accesses from an FPGA fabric to the rest of the system. Also, assume that the attacker cannot be in physical proximity of the target device. In this scenario, an attacker can instantiate ring oscillators (ROs), or a delay line coupled with time-to-digital converters (TDCs) to measure the voltage fluctuations on the shared power distribution network (PDN) that are caused by the target circuit [4-6]. Because an FPGA is programmed by loading a bit-stream in software, an adversary who has access to bit-stream data, or has permission to program at least a part of an FPGA, can perform power side-channel attacks remotely. This attack vector can also be applicable to discrete FPGA architecture that shares the same power supply with a CPU on the system, or data-centres that allows multiple users to co-share the FPGA resources [7,8].

For heterogeneous SoC-FPGA architectures, it is common for modules on the same die to share the same power supply. For example, in Xilinx Zynq SoC both CPU and FPGA fabric share the same PDN. The PDN converts and distributes power from the power supply to individual circuit components (in this case CPU and FPGA fabric), with a goal to provide a clean voltage supply resistant to varying current demands [3,4].

To maintain a constant voltage, a PDN uses a voltage regulator to adjust the amount of supplied current and uses decoupling capacitors as a buffer to handle current variations. However, the voltage regulator and the decoupling capacitors cannot completely hide current variations, and high switching activities often lead to transient voltage drops in the PDN of an FPGA [3]. The voltage drop on the PDN reflects the power consumption [3]. Additionally, a change in the combinational logic delay reflects the voltage drop that correlates with the power consumption and the switching activity of the circuits [3]. This correlation between combinational logic delay and the power consumption can be leveraged to build an on-chip power monitor that will allows us to measure a combinational path delay, and further estimate the power consumption of other modules that share the PDN [3].

The FPGA based on-chip power monitors are typically either TDCs or ROs. Both circuits exploit propagation delay as a proxy for measuring supply voltage, as lower supply voltage is known to cause an increase in the propagation delay [9]. TDCs detect voltage changes in the FPGA PDN by sensing changes in the delay of a propagating signal through a chain of buffers or other logic [9,12]. TDC sensors can also be used as receivers for covert communication from information leaking hardware Trojans in the target circuit [9,12]. ROs based sensors can also be used to monitor the supply voltage of an FPGA PDN because the propagation delay through the RO, which depends on supply voltage, can be observed by measuring oscillation frequency [10,11].  

Example Approaches: • Identify different power monitors and compare with existing ROs and TDCs power monitors.
• Investigate different ROs designs for power monitoring and estimate their entropies.
• Validate the capability of remote power side-channel attacks against modern FPGA architectures.
• Develop countermeasures against remote power side channel attacks.

Unclassified Relevance to the Intelligence Community:
This research will inform understanding and confidence in FPGA based SoC designs, characterise potential remote power side channel attacks and inform potential mitigation strategies.

References:
1. Trimberger and McNeil, “Security of FPGAs in data centers,” in 2017 IEEE 2nd International Verification and Security Workshop (IVSW), 2017
2. Gnad, et al. “Voltage drop-based fault attacks on FPGAs using valid bitstreams,” in 27th International Conference on Field Programmable Logic and Applications (FPL), 2017.
3. Pant, “Design and analysis of power distribution networks in VLSI circuits,” Ph.D. dissertation, The University of Michigan, 2008.
4. Masle and Luk, “Detecting power attacks on reconfigurable hardware,” in 22nd International Conference on Field Programmable Logic and Applications (FPL), 2012.
5. Barbareschi et.al., Implementation and Analysis of Ring Oscillator Circuits on Xilinx FPGAs. Springer International Publishing, 2017.
6. Hoque, “Ring oscillator based hardware trojan detection,” Master’s thesis, University of Toledo, Toledo, Ohio, USA, 2015.
7. Amazon EC2 F1, https://aws.amazon.com/ec2/instance-types/f1/ , Amazon.com, Inc, accessed: 2017.
8. Chen et al. “Enabling FPGAs in the cloud,” in 11th ACM Conference on Computing Frontiers (CF), 2014.
9. Dennis et al. 2016. Analysis of transient voltage fluctuations in FPGAs. In International Conference on Field-Programmable Technology.
10. Zhao and Suh. 2018. FPGA-based remote power side-channel attacks. In 2018 IEEE Symposium on Security and Privacy (SP’18). IEEE.
11. Kenneth and Hayes. 2012. Low-cost sensing with ring oscillator arrays for healthier reconfigurable systems. ACM Transactions on Reconfigurable Technology and Systems 5, 1 (2012).
12. Kenneth and French. 2013. Sensing nanosecond-scale voltage attacks and natural transients in FPGAs. In ACM/SIGDA International Symposium on Field Programmable Gate Arrays.

ORISE Topic Number: ICPD-2026-50-Warfare

Title: The impact of artificial intelligence and machine learning on chemical and biological warfare

Keywords: Artificial intelligence, drug design, pharmacology, machine learning, neural networks, algorithm, MegaSyn, BioNavi, Chemistry42, proteomics, computational biology, synthetic biology, chemical weapons, biological weapons

Description: Artificial intelligence (AI) enabled tools are expanding the scope of chemical and biological sciences. Previously unknown molecules and organisms are now being discovered and previously unreachable molecules are becoming more accessible. AI enabled tools are being deployed in the pharmaceutical industry, biotechnology and genetic engineering, consumer product manufacturing (cosmetics, avoidance of animal testing and to reduce environmental impacts), and in agriculture (biodegradation and reducing toxicity to non-target species). This broad-ranging scientific trend may also be used to design or produce new chemical and biological structures similar or superior to known chemical and biological warfare (CBW) agents. The potential capacity for rapid growth of available CBW threats with enhanced or comparable lethality to presently known toxic agents could potentially overwhelm international arms controls and countermeasure capability and will require the proactive development of new surveillance methods and novel detection, identification and mitigation systems.

Applicants should approach the topic with intent of undertaking a literature review and feasibility analysis of predictive AI algorithm applications and trends with respect to potential impacts on chemical and biological warfare controls.  

Example Approaches: Research proposals could approach this issue from a variety of disciplines, or as a cross-disciplinary effort. The problem touches on aspects of chemistry, biotechnology, synthetic biology, computer/software engineering, neural networks and machine learning (ML), applied science, innovation policy, and pharmacology. Proposals may consider ways to monitor and mitigate threats of generatively designed agents using:
• machine learning models that use publicly available datasets and open-source generative software
• additional machine learning tools to model parameters such as environmental and metabolic stability
• retrosynthesis tools (commercially available or open-source)
• identification of suitable “chemistry/biology starting points”

Unclassified Relevance to the Intelligence Community:
Artificial intelligence-based generative design is a disruptive (emergent and convergent) technology with the potential to generate new threat agents, or increase the threat from current agents. The combination of technological advances in this field, coupled with the limited regulations associated with rapidly developing AI technologies, may result in proliferation of AI-enhanced threats. A deeper understanding of the latest technological improvements in the AI and ML fields, the potential applications of AI/ML in a CBW context, and the prospects of yechnologies to protect/defend against these applications, are critical to informing warnings and indicators for the intelligence community.

Information relevance to the intelligence community could be prioritised as follows (according to technology maturity):
1. Safeguard implications for artificial intelligence enhanced CBW agent design. The dual-use nature of artificial generative design technology advancements and potential impacts on chemical and biological threats.
2. Safeguard implications for technological/design hurdles from theoretical to practical design outputs – including; current limitations of AI/ML systems, overcoming predictive failure and computational structure validation challenges, and future impacts of quantum machine learning.
3. Safeguard implications of adaptive AI/ML to develop strategies and materials to proactively avoid controls, detection and countermeasures.

Understanding these emerging technological possibilities will lay the knowledge foundations to provide the NIC with the insight necessary to address risks associated with the field in terms of national security and global proliferation.

References:
De Lima RC, Sinclair L, Megger R, Maciel MAG, Vasconcelos PFDC, Quaresma JAS. Artificial intelligence challenges in the face of biological threats: emerging catastrophic risks for public health. Front Artif Intell. 2024 May 10;7:1382356. doi: 10.3389/frai.2024.1382356.

ORISE Topic Number: ICPD-2026-51-Rapid

Title: Rapid detection and identification of environmental pathogens

Keywords: Environmental pathogen, genome, DNA sequence, real-time sequencing, diagnostics, biosurveillance

Description: Environmental pathogens are estimated to cost Australia billions of dollars every year, through causing human disease, endangering native animal and plant species, or damaging the environment and agricultural ecosystems. Plant pathogens pose significant threats to agricultural output, while human pathogens can cost millions more in financial losses in healthcare and lost worker productivity. Their effects include lost production, environmental damage, and the endangerment of native species. Some pathogens can also cost millions more in healthcare and lost productivity.

Microbial pathogens (bacteria, viruses and some fungi) cannot be readily detected and accurately identified in the environment, at the point of exposure. In these cases, samples are sent to a laboratory for analysis – a relatively slow and inefficient process that may yield inconclusive results depending on sample quality, sample preparation and handling, and technological post processing.

Identifying microbial or viral environmental pathogens in a way that is rapid, accurate and sensitive remains a challenge. Identification would likely require a broad-spectrum assay that includes the ability to identify unique genetic signatures for accurate strain identification. However, this capability would have broad applicability for bio-surveillance and safety in various industries such as agriculture, food safety and healthcare.

This project aims to explore state-of-the-art methods, so that control and oversight authorities are equipped to detect and accurately identify pathogenic microorganisms in the environment. Future technology should be adjustable to allow new species to be added to the control pool in order to keep up with the rapid evolution (or development i.e. through genetic engineering) of new species of microbes and viruses of concern.

Applicants should approach the topic with intent of undertaking a literature review and feasibility analysis of current and emerging technologies and processes for the detection and identification of environmental pathogens that inhabit all domains (air, water and land). Then, through research and analysis, determine how such technologies, equipment or processes can be optimised for rapid and accurate detection at the point of exposure to environmental pathogens.  

Example Approaches: Research proposals could approach this issue from a variety of disciplines, or as a cross-disciplinary effort. The problem touches on aspects of biotechnology, genomics, synthetic biology and engineering. The proposal should focus on exploring new platform capability for the detection and identification of a range of pathogenic microorganisms, but could initially focus on the detection and identification of a single pathogen with high selectivity and sensitivity with the potential to broaden the platform in the future. Investigations could include a range of novel methodologies.

Unclassified Relevance to the Intelligence Community:
The national security implications of environmental pathogen detection and identification are significant. At the national level, insight into such emerging technologies would better enable the monitor and report on the spread of human or crop diseases. Detection of pathogens is critical for strategic and operational awareness; and for assigning accountability under international treaties, peace agreements and global norms.

References:
Xu, J., M. Akhtar, W. Meng, J. Bai, S. Prince, and R. Huang. 2025. “ Advances in Pathogen Detection: From Traditional Methods to Nanotechnology, Biosensing and AI Integration.” Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology 17, no. 4: e70022. https://doi.org/10.1002/wnan.70022.

ORISE Topic Number: ICPD-2026-52-Microwaves

Title: Multifunctional metamaterials for broadband absorption of high-power microwaves

Keywords: Multifunctional metamaterials; multi-physics simulations; electromagnetic absorption; system reliability; multi-material additive manufacturing

Description: Emerging development and deployment of high-power microwave (HPM) weapons (Benford 2024) will likely increase the vulnerability of satellites, aircraft and terrestrial devices. Electromagnetic metamaterials, as artificially engineered structures, can be tailored to either absorb or reflect HPM, depending on their material compositions and structural configuration, which offer a promising route for improved device protection. Nevertheless, conventional electromagnetic metamaterials only exhibit narrowband protection and generally fail to address the coupled electromagnetic–thermal–mechanical interactions that occur under broadband HPM exposure. Such coupling effects can lead to localised overheating, material degradation, and mechanical instability, which in turn trigger cascading failures that pose threats to device survivability and reliability.

This interdisciplinary topic aims to develop multifunctional metamaterials (Li et al. 2022) that exhibit broadband electromagnetic performance, efficient thermal management, and enhanced mechanical stability to minimize the adverse effects of HPM on device survivability and reliability in a range of environments. The project would explore a new generative design approach that systematically integrates multi-physics simulations, advanced system reliability assessment methods, rational design optimisation approaches, and multi-materials additive manufacturing techniques to design multifunctional metamaterials for improving the survivability and reliability of devices against HPM threats.  

Example Approaches: • Conducting multi-physics simulations to investigate the coupled electromagnetic–thermal–mechanical behaviours of multifunctional metamaterials.
• Developing advanced rational design optimisation approaches to generate multifunctional metamaterials with broadband microwave absorption, efficient thermal management, and enhanced mechanical stability.
• Utilizing multi-material additive manufacturing techniques for prototyping and conducting experimental characterization of the electromagnetic, thermal, and mechanical properties of meta-lattices under HPM excitation.

Unclassified Relevance to the intelligence community:
This project would explore avenues for advanced materials technologies that may address emerging HPM threats to electronic devices. The project will experiment with multifunctional metamaterial systems capable of protecting devices against HPM threats, to enhance survivability and reliability of critical assets.

References:
• Benford J. History and future of high power microwaves[J]. IEEE Transactions on Plasma Science, 2024, 52(4): 1137-1144.
• Li Z, Gao W, Wang M Y, et al. Design of multi-material isotropic auxetic microlattices with zero thermal expansion[J]. Materials & Design, 2022, 222: 111051.
• Yang Y, Yin Z, Zhu X, et al. A review of multimaterial additively manufactured electronics and 4-D printing/origami shape-memory devices: Design, fabrication, and implementation[J]. Proceedings of the IEEE, 2024.

ORISE Topic Number: ICPD-2026-53-Neuromorphic

Title: Visual language model that will be able to leverage neuromorphic and quantum for HSI

Keywords: large language model, neuromorphic computing, Quantum computing, hyperspectral image (HSI)

Description: (U) Hyperspectral image understanding for foundation data generation is a challenging problem for earth obseration mission, current method usually only work for electro-optics images, and less efficient for high dimension data. We are looking for non-traditional computing approach that out-perform traditional GPU computing, which requires innovation on tokenization scheme.

Example Approaches: Neuromorphic computing is one example approach that leverage spiking nerual network and amplitude encoding. We will leverage feature learning and use traditional rate model as teacher in the training process so that we can develop spiking neural network that can be deployed to neuromorphic chip. Meanwhile, quantum computing is another approach that we can explore using quantum circuit, which can leverage quantum computer's capability.

ORISE Topic Number: ICPD-2026-54-Nuclear

Title: Atomic Nuclear and optimal clock

Keywords: Quantum Sensor; Integrated Photonics, Atomic Sensors; Ion; Nuclear/Atomic Clock, Optical Clock

Description: Optical clocks can outperform traditional microwave clocks by factors of 100 – 1000 in stability and have become some of the most precise measurement devices ever built. However, current precision optical clocks are mostly laboratory-sized. While recent work by multiple groups around the world attempts to reduce the size of fieldable systems, the concepts being pursued are still trailer or full-instrument rack in scale, at best. Significant miniaturization is necessary to achieve scalable, portable devices needed for these applications. Recent demonstrations have shown that ion optical clocks are particularly well suited to an integrated platform. Recent advances in research and development of nuclear based timekeeping devices have also sparked interest in new applications of these types of clocks. Work related to this new technology is encouraged.  

Example Approaches: Research for this topic may include a variety of methods to demonstrate a path to a fully integrated clock.

Unclassified Relevance to the Intelligence Community:
Timing holdover in the absence of GPS, is a serious challenge. One way the DoD/IC is looking to address this is via portable and ultra-stable clocks. We aim to develop a clock capable of holding ns stability for month duration. The envisioned sensor will provide quality data to a resolution capable of supporting DOD’s Alternative Positioning and Navigation applications.

Unclassified Linkage to current DNI's S&T Priorities:
2N098 Develop/enhance video and electro-optical processing capabilities that integrate computational analytic methods 2N002 Develop/enhance capabilities to collect information on global science and technology activities
2N061 Develop/enhance capabilities to identify and protect critical assets, information systems, technologies, industries, and people.   

ORISE Topic Number: ICPD-2026-55-Community

Title: Artificial Intelligence Evaluation for Intelligence Community Uses

Keywords: Artificial Intelligence, Large Language Models, Biological Design Tools, Membership Inference Attacks, Unlearning, Human-Machine Team, AI Evaluation, Mechanistic Interpretability

Description: Artificial Intelligence (AI) presents significant opportunities for the U.S. Intelligence Community (IC) to advance mission effectiveness, but realizing these gains requires rigorous, IC-specific evaluation. This research topic focuses on advancing the design and science of AI evaluations, benchmarks, and metrics to assess (1) technical performance of cutting-edge AI models across IC use cases, (2) adversary AI capabilities to reduce strategic surprise, (3) information security and biosecurity risks from large language models (LLMs) and biological design tools (BDTs), and (4) integrated human-machine teaming (HMT) and AI-augmented decision-making in a range of conditions. The research should aim to produce benchmarks and empirical studies, characterize risks and limitations in existing and emerging AI evaluation methods, and accelerate responsible AI adoption in the U.S. Intelligence Community.

Example Approaches: - Characterize key trends, patterns, risks, limitations, and sources of legitimate disagreement or variation across AI evaluation methods. Develop methods for monitoring and updating consensus and alternative evaluation methods for AI systems across various communities and use cases.
- Develop empirically derived evaluation requirements for deployment of AI in safetycritical IC missions (See Morey, D.A., Rayo, M.F. & Woods, D.D. Empirically derived evaluation requirements for responsible deployments of AI in safety-critical settings. npj Digit. Med. 8, 374 (2025). https://doi.org/10.1038/s41746-025-01784-y)
- Develop safety protocols and audit mechanisms for autonomous and semi-autonomous AI agents in a variety of contexts
- Assess processing error logs to identify and address systematic failure causes in image processing computer vision algorithms. Develop methods to analyze vast amount of processing runs to identify system limitations.
- Advance the understanding of risks and opportunities associated with unlearning by developing scalable, provable, and adaptable methods for life science LLMs and BDTs.
- Develop processes, pipelines, and metrics to effectively and automatically baseline the capabilities of newly released LLMs and BDTs. Assessments should focus on potential for misuse by malicious actors, and the emergence of novel, revolutionary capabilities not present in existing models.
- Advance methods to decode the internal logic of BDTs and frontier models to ensure that researchers can distinguish between valid biological reasoning and dangerous dataset artifacts.
- Develop membership inference and model inversion techniques specifically tailored to biological design tools (BDTs). Build capabilities to analyze data leakage in biological sequence and structural datasets used for model training.

ORISE Topic Number: ICPD-2026-56-Military

Title: Adaption of Advanced Civilian Diagnostic medical Devices for Military Environments

Keywords: Analysis, Molecular Diagnostics, Wearable Sensors, Telemedicine, Ruggedization, Data Security, Human Factors Engineering

Description: The civilian sector has seen a surge in advanced, portable diagnostic medical technologies, offering the potential for rapid and accurate assessments at the point of care. Military medical personnel could greatly benefit from such innovations, but the devices often lack the ruggedness and adaptability needed for deployment in challenging operational environments. This research focuses on identifying, evaluating, and adapting promising civilian diagnostic technologies for military use, ensuring their reliability and effectiveness in austere settings.
Military medical care demands rapid and accurate diagnostics, often in resource-limited and environmentally harsh conditions. While some military-specific diagnostic tools exist, the pace of innovation in the civilian medical device industry presents an opportunity to enhance battlefield medical capabilities. This research addresses the challenges of transitioning specific diagnostic technologies, including but not limited to:

Point-of-Care Ultrasound (POCUS): Handheld ultrasound devices offer non-invasive imaging for rapid assessment of injuries and internal conditions.
Portable Blood Analyzers: Compact devices for rapid blood tests, providing critical information on electrolyte balance, blood cell counts, and other vital parameters.
Molecular Diagnostic Devices: Devices for detecting infectious diseases, biomarkers, and genetic signatures using PCR or other molecular techniques.
Wearable Physiological Monitors: Sensors that continuously monitor vital signs, such as heart rate, respiration rate, and blood oxygen saturation.

These technologies, while readily available in civilian settings, require significant adaptation to withstand the rigors of military deployment. Issues such as durability, power requirements, data security, environmental tolerances, and usability under stress must be addressed.
The central focus of this postdoctoral research is to develop strategies for ruggedizing, adapting, and validating these civilian diagnostic devices for military use. This includes exploring methods for improving their physical resilience, optimizing power consumption for extended operation, ensuring data security and interoperability with military communication systems, and adapting device interfaces for ease of use by medical personnel in stressful situations.

Example Approaches: Material Science for Ruggedization of Diagnostic Devices: Focus on finding rugged and resilient materials with a wide range of protection.

Low-Power Optimization for Diagnostic Operations: Focus on improving efficiency of diagnostic tools through software and hardware integration.

Wireless Integration for Telemedicine Applications: Focus on incorporating wireless modules to transmit medical data. Data Security and Encryption Protocols for Diagnostic Transmissions: Focus on encrypting and securing transmitted data. Usability Assessment: Conduct trials to determine how best to implement these tools.

Standardization and Interoperability: Focus on how best to integrate these tools with pre-existing infrastructure.

ORISE Topic Number: ICPD-2026-57-Bioremediation

Title: Advancements in bioremediation of energetics and explosives

Keywords: energetics, bioremediation, mines, unexploded ordnance, bioengineering, synthetic biology, Purdue, GTRI, Caltech, MIT,

Description: Traditional bioremediation efforts targeting energetic compounds—such as TNT, RDX, and HMX—and their associated production chemicals have primarily focused on contaminated soils at legacy munition sites, former military installations, and underwater ordnance disposal areas. These efforts have demonstrated promising results using microbial consortia, enzymatic degradation, and phytoremediation techniques to reduce environmental and human health risks posed by residual contamination.

This research seeks to extend the frontier of bioremediation by exploring its application to dispersed energetic materials, including land mines and unexploded ordnance (UXO). Unlike conventional environmental contamination, UXOs and mines presents unique challenges due to the potential for artificial barriers, metal and plastics, along with unpredictable distribution in terrestrial and aquatic environments. The goal is to investigate biologically driven methods that can gradually yet effectively neutralize or destabilize energetic compounds, even wrapped to protect against environmental exposure, within these devices, safely rendering them inert over extended timeframes - days, weeks, or even years.  

Example Approaches: 
-Discovery and characterization of native consortia and enzymes which can more rapidly transform energetic residues and/or their encapsulates.
-Engineering of safe and controllable microbial systems and enzymes which can more rapidly transform energetic residues and/or their encapsulates.
-Development of stable, non-propagating biocatalyst formulations and delivery concepts that improve bioavailability while avoid mechanical or physical disturbance of systems.
-Methods for improving detection and removal, specifically aligned with features which fall within the visible light spectrum.

ORISE Topic Number: ICPD-2026-58-DigitalCyber

Title: Advancing physical and digital cyber biosecurity for instruments and facilities

Keywords: digital biosecurity, cyber biosecurity, biotechnology, instruments, facilities, laboratories, factories, biomanufacturing

Description: As biological instrumentation evolves, the convergence of physical systems with digital networks is reshaping the operational landscape of bio-research environments. Instruments are no longer isolated tools; they are increasingly embedded within interconnected ecosystems that span
diverse facility types—from factories to high-containment laboratories to farms. This integration introduces novel vectors for exploitation, particularly where biologically relevant data, control systems, and operational protocols intersect with digital infrastructure. The complexity and heterogeneity of these deployments demand a reassessment of biosecurity paradigms, especially as traditional safeguards may not adequately address emerging threats posed by remote access, autonomous operation, or cross-platform interoperability.

This research initiative aims to rigorously evaluate and mitigate biosecurity vulnerabilities across both physical and digital domains, with a specific focus on the hardware elements, or data, which are linked to interacting or measuring the biological material itself and not found in any other component hardware outside of biology. Components found across other non biology-related instruments are not in scope for this effort.  

Example Approaches: 1. Securing Biotechnologies from Stand-off Exploitation: Investigate the risks and identify methods to prevent exploitation (e.g., acoustic, electromagnetic, thermal, or optical signatures) of the unique hardware components associated only with biological instruments.
2. Methods for reducing noise from externally sources: Identify risks and methods for reducing external stimuli from creating uncontrollable noise or data issues associated with the hardware or facility components unique to biotechnology and biology.
3. Securing Biotechnologies and Biotechnology Infrastructure from Real-Time and Delayed Exploits: Explore the risks and mitigation opportunities for digital attack surface unique to biological instruments, inclusive of the possibility of man-in-the-middle (MitM), firmware-level, and hardware-based exploits.
4. Secure Handling and Encryption of Biological Data: Develop methods for the secure storage, transmission, and processing of biological data—whether raw genomic sequences, assembled datasets, or metadata—within instruments and facilities.

ORISE Topic Number: ICPD-2026-59-Sweltering

Title: SWELTERING YETI: AI Optimized Extreme Temperature Electrolytes

Keywords: Artificial Intelligence, Energy and Power, Applied AI/ML, Electrolyte Optimization, Extreme
Temperature Performance

Description: The Intelligence Community (IC) often requires batteries that perform in niche extreme temperature environments not representative of typical consumer use, creating a technological gap between IC needs and industy offerings. Batteries are currently manufactured to operate in a wide temperature window (~−10 °C to 45 °C) with baseline operation at 24 °C because most consumers use devices under standard conditions. Manufacturers that do target low/high temperature (<−10 °C or >45 °C) performance generally make only minor cell-design or electrolyte-formulation changes that can improve specific metrics (shelf-life, power delivery, cycle life, capacity) in limited regimes, but rarely all simultaneously. As a result, performance at truly extreme temperatures (<-30°C
or >60°C) is typically poor or nonexistent.

Electrolyte development has historically relied on iterative, resource-intensive testing of solvents, salts, and additives, with focused effort on regimes most relevant to commercial consumers. Components or formulations that might enable extreme-temperature performance are often discarded early or left unexplored. Modeling of electrolyte formulations is challenging because of large, multi-variable dependence, so models have generally only informed limited portions of the test matrix.

Targeting extreme temperature operational windows (−60 °C to −10 °C and 45 °C to 90 °C) to improve electrolyte stability and battery performance, is underexplored. Developing electrolytes that enable discharge performance at these extremes while maintaining shelf-life stability at room temperature during shipping/handling and storage would open use of high-energy-density power sources in currently inaccessible locales.

Example Approaches: AI/ML tools are uniquely poised to explore the expansive chemical space of novel electrolyte formulations and cell designs efficiently. These tools can systematically search and optimize regions of composition space that would be slow and cost-prohibitive to explore with a purely Edisonian approach.

Use of AI/ML toward electrolyte optimization should aim to identify components and combinations that enhance stability and efficiency in extreme temperature regimes while reducing the time and cost of iterative testing and validation. Opportunities exist across a variety of battery chemistries, including Li-ion rechargeables (NMC/Graphite, NMC/Silicon, LFP), primary chemistries (CFx, MnO2), and next-generation chemistries (Li-metal). Solid-state electrolyte formulations are not of interest.

We recognize that small-scale test cells (e.g., coin-cells) and literature data may not fully capture the thermal, mechanical, and interfacial behavior of industry-scale cell formats (e.g., pouch cells) and that Artificial intelligence (AI) and machine learning (ML) models trained only on research-scale data risk limited transferability. Therefore, it is preferred that proposals include a clear plan to close form-factor gaps, such as training the model on pouch-cell data or pretraining on coin-cell data then fine-tuning and validating on pouch-cells.

We encourage submissions that balance rapid discovery in bench-scale cells with deliberate strategies to accelerate translation to industry-relevant formats and operating conditions under either extreme cold or hot temperatures. Alternatively, an optimized AI/ML platform that can accept inputs of cathode/anode chemistry, cell dimensions, voltage operating window, temperature window, and expected shelf-life variables towards electrolyte optimization would also be of interest

ORISE Topic Number: ICPD-2026-60-lowresource

Title: Identity intelligence from low-resource image and video with unstructured multimodal auxillary data

Keywords: identitiy resolution, image analysis, video analysis, multimodal data fusion, knowledge graphs, artificial intelligence, open source intelligence (OSINT)

Description: The National Digital Exploitation & OSINT Center is seeking exceptional candidates for a Postdoctoral Research Fellow position focused on advanced image and video processing, specifically on developing, refining and deploying state-of-the-art and beyond state-of-the-art systems that combine low-quality, low-resource image and video information with other data (e.g. knowledge graphs, open source intelligence, domain-specific knowledge, etc.) to identify individuals or groups with higher accuracy than is possible today. This unique opportunity allows early-career scientist s to gain experience on problems of national significance and deploy research prototypes into operational environments. This position also allows for collaboration with other members of the intelligence community as well as US national labs and the National Institute of Standards and Technology (NIST). This position offers a valuable stepping-stone for academic, government and industry career paths.

We are particularly interested in candidates with solid expertise in:
• Biometric pipelines based on still frames and video.
• Approaches for improving data quality.
• Non-traditional biometrics, such as gait analysis or iris identification at a distance.
• Multi-modal models that combine deep learning with symbolic representations.
• Reinforcement learning for parameter tuning.
• Model analysis, especially approaches for automated model analysis.
• Edge solutions, and computing in low-resource environments.
• Embodied AI, robotic AI, human-AI collaboration

Example Approaches: Evaluate state-of-the-art identity intelligence algorithms integrating artificial intelligence capabilities and compare them to algorithms that are currently deployed on operational systems.
- Explore opportunities for creating principled, experiment-conducting, automated pipelines to allow for repeated experiments at scale.
- Create multi-modal pipelines and operational systems that combine imagery, video, metadata, and other data sources to improve and validate identity intelligence information.
- Engage with contractors to implement experimental algorithms in an operational environment.
- Publish high-impact articles in both government and academic venues

ORISE Topic Number: ICPD-2026-61-Biosurveillance

Title: Genetic Forensics in Bio-surveillance

Keywords: Genetic Forensics, Bioinformatics, Bio-surveillance, Synthetic Biology, Microbial Genomics

Description: This research initiative is focused on developing next-generation bioinformatics tools and curated databases to advance the field of environmental bio-surveillance. By capitalizing on breakthroughs in CRISPR, next-generation sequencing, and synthetic biology, the project aims to detect and characterize previously uncatalogued genetic variations, including mutations, modifications, and edits, within ubiquitous microbial populations. Moving beyond the constraints of current methods that rely on alignment to established reference genomes, this project will establish a new analytical framework to systematically identify novel genetic features from samples and link them to their functional and ecological roles, providing unprecedented insight into microbial adaptation and evolution. The developed analytical framework will be modular and adaptable with an emphasis on incorporation into non-target sequencing methodologies.

The central problem this research addresses is the need for more analytically robust bioinformatic tools to accompany new and novel sequencing methodologies. This research emphasizes the establishment of novel tools that provide rapid whole genome analytical characterization and allow for detection and differentiation of emergent mutations from wild type variants. This effort is intended to enhance the resolution of data collected via sequencing and provide more back-end analytical capability.    

Example Approaches: 

• Database of Genetic Variations in Ubiquitous Microbes: One key research direction could involve developing a comprehensive database that catalogues genetic variations in widely distributed organisms. This could include naturally occurring mutations or modified genes that influence traits such as antibiotic resistance, metabolic pathways, or environmental adaptability.
• Detection of Modified or Mutated Genes: Postdocs could focus on developing advanced bioinformatics methods for detecting genetic modifications that occur in common organisms, whether naturally or through environmental pressures. This research could help reveal the genetic plasticity of microbial communities and provide insight into how microbes evolve in response to changing conditions.
• Functional Genomics and Ecological Implications: Research could investigate how these genetic variations contribute to ecological roles, such as nutrient cycling, symbiosis, or pathogen resistance. By linking genetic data to ecological functions, this project could uncover novel microbial traits with implications for environmental science.
• Machine Learning for Predicting Ecological Impact: Machine learning models could be developed to predict the ecological consequences of genetic variations. These models would help assess how mutations or gene edits in ubiquitous microbes affect their behavior and their interactions with the environment, offering new tools for ecological discovery and monitoring.
• Integration with Environmental Monitoring Systems: Research could also explore how new analytical tools can be integrated with environmental monitoring systems, enabling more effective and real-time analysis of microbial populations in natural ecosystems