University of Virginia Main Campus

NSF Awards · FY 2026 · 2026-10

This project studies how to create synthetic datasets that retain useful patterns from sensitive data while protecting privacy of individuals. Many hospitals, companies, public agencies, and researchers need data to improve services, test ideas, etc., but they often cannot share original records because they contain private information. This project addresses this gap by making data sharing safer and more useful. The project's novelties are creating a general way to break synthetic data generation into two connected steps, new methods that combine classical statistical ideas with modern learning tools, and systematic ways to use public data and existing models without weakening privacy protection. The project's broader significance and importance are that it expands safe access to data for research and education, strengthens privacy practice in data-driven fields, and creates training and research opportunities for students. Specifically, the research develops a framework that separates synthetic data generation into information extraction from sensitive data under formal privacy protection based on differential privacy and reconstruction of synthetic data from the extracted information. Within this framework, the project has three research thrusts. First, for tabular data, it examines why statistical methods often outperform neural network methods and designs hybrid methods that combine strengths from both approaches. Second, for image and multimodal data, it studies adaptive high-order projections, including Fourier representations, to capture broad structure and preserve relationships across data types. Third, it develops a double-cone framework for selecting, expanding, and adapting public data sources and for using existing models so that public information can improve synthetic data quality in a systematic way. The project also brings these ideas into courses, student research, open-source tools, and public demonstrations. The expected results are stronger foundations and more practical methods for privacy-protected synthetic data generation across application areas. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-10

Knowledge acquisition—the ability of artificial intelligence (AI) systems to extract actionable insights from vast amounts of unstructured text—is critical for advancements in healthcare, education, and scientific discovery. While Large Language Models (LLMs) have shown impressive capabilities, their reliability depends heavily on massive, perfectly curated datasets, which are expensive and often unavailable in specialized domains. This CAREER project addresses this bottleneck by developing a new paradigm called “structure-aware weak supervision.” Instead of relying on perfect human annotations, the project enables AI systems to learn autonomously from incomplete, noisy, and ambiguous data by discovering and utilizing underlying semantic structures, such as concept hierarchies and retrieval pathways. By reducing the dependency on expensive labeled data, this research democratizes the development of highly accurate, domain-specific AI tools for resource-constrained environments, such as public health agencies and community organizations. The project also integrates these research outcomes into new undergraduate and graduate curricula, open-source educational toolkits, and targeted K-12 outreach programs designed to broaden participation in computing and teach the next generation how to build reliable, human-centered AI systems. This project proposes a unified framework for learning under weak supervision by bridging unstructured language data with structured, interpretable knowledge representations. The research is organized into three synergistic thrusts. Thrust 1 tackles incomplete supervision by inducing latent ontologies from unlabeled corpora via a novel Spherical Hierarchical Expectation-Maximization (SHEM) algorithm, enabling scalable information extraction and classification without predefined schemas. Thrust 2 addresses noisy supervision by designing a Denoising Retrieval-Augmented Generation (DeRAG) framework. It integrates symbolic reasoning over the induced ontologies with Structure-Aware Contrastive Retrieval (SACRet) to actively filter distractors and reliably ground language model outputs. Thrust 3 tackles ambiguous supervision by modeling complex, multi-faceted human preferences. It introduces a Tree of Reward Models (TreeRM) and Hierarchical Dirichlet Thompson Sampling (HDTS) to capture both shared foundational values (e.g., safety, factuality) and personalized user preferences (e.g., tone), ensuring robust AI alignment. Together, these contributions advance the theoretical foundations and practical methodologies of knowledge-centric AI, creating systems that autonomously construct knowledge, dynamically adapt to supervision gaps, and reliably align with hierarchical human values. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-08

Computer simulations, based on physics, can be used to predict how galaxies grow over time. These cosmological simulations connect the growth of dark matter to stellar and gaseous processes. The predictions that these simulations provide are often sensitive to various arbitrary choices which introduce uncertainty that is rarely quantified. This program will develop a novel analysis, based on AI, that will allow for the characterization of uncertainties in cosmological simulations in a novel way. This research will be conducted at the University of Virginia and support a graduate research student who will participate in the model development. Educational initiatives include a Research Experiences for Undergraduates program and an annual schedule of teacher training activities. This project will advance the field of galaxy formation by executing a simulation suite designed for uncertainty quantification across physical models, cosmological assumptions, and resolution scales. A fully automated simulation pipeline will enable the execution of thousands of zoom-in simulations at varying resolution and halo mass, sampling parameters such as supernova feedback strength, AGN efficiency, Omega-matter, and sigma-8. Simulation outputs will be paired with dark matter-only merger trees and analyzed using Graph Neural Networks (GNNs) that learn to predict galaxy properties across the sampled parameter space. The models will be extended to predict spatially resolved galaxy properties using a hybrid of GNNs and diffusion-based generative models, allowing the sampling of entire profile distributions conditioned on both formation history and simulation assumptions. The tools will enable direct comparison across physical models and offer a statistically grounded framework for testing small-scale tensions in Lambda-CDM. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-07

Infectious disease outbreaks pose serious threats to global health, economic stability, and societal well-being. An effective response system needs to answer the following key questions in a timely manner: how will a disease spread, what interventions can control them, and how to monitor populations to enable early warnings? However, current approaches often rely on fragmented or delayed data, such as clinical case reports, incomplete testing coverage, and/or the inability to see how infected people move and interact. These factors make it difficult to combine different sources of information and adapt to rapidly changing conditions. These limitations can delay response efforts and reduce their effectiveness. This project aims to develop next-generation epidemic intelligence systems that improve how public health agencies forecast, manage, and monitor infectious diseases. By enabling earlier detection, more targeted interventions, and better situational awareness, the project will strengthen public health infrastructure, support informed decision-making, and enhance resilience to future outbreaks. The project also contributes to education by training students at multiple levels, engaging K-12 learners, and providing open resources to increase participation in data science and public health. To meet these goals, this project develops a unified, data-driven framework that integrates epidemic forecasting, intervention planning, and surveillance under diverse and evolving data conditions. Specifically, the project focuses on three objectives: (1) advancing epidemic forecasting by combining established transmission models with modern machine learning to enable multi-scale prediction across fine-grained contact networks and large-scale population dynamics; (2) designing context-aware intervention strategies that adapt to local conditions, such as population density, regional transmission risk, and operational constraints, while balancing effectiveness and cost; and, (3) developing resource-efficient surveillance methods that optimize diagnostic testing and data collection across multiple sources, including clinical reports and wastewater signals. To address changing epidemic conditions, the framework incorporates adaptive learning mechanisms that detect shifts in data patterns and update models and policies accordingly. The project will be evaluated using real-world datasets and simulation environments, and the resulting methods will provide scalable and robust tools for epidemic analysis, contributing broadly to the fields of data science, network modeling, and public health. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-06

The United States faces a growing gap between demand for artificial intelligence (AI) literacy and the capacity of K–12 education to deliver it. Despite the rapid expansion of AI across virtually every industry sector, most high school educators lack the training to teach AI concepts, and most students graduate without meaningful exposure to this transformative technology. This project addresses that gap by integrating AI education into high school classrooms, empowering teachers, engaging students in hands-on learning, and connecting students to real workforce opportunities. By doing so, the project advances national priorities articulated in Executive Order 14277, "Advancing Artificial Intelligence Education for American Youth," and contributes to U.S. competitiveness in the global AI economy. The project employs a train-the-trainer model organized around four objectives. First, twenty high school teachers will be trained over three years through "AI on Saturdays," a 14-week professional development program in Python programming, machine learning, and AI application development, followed by curriculum design studios that produce standards-aligned, adaptable instructional modules. Second, participating teachers will deliver a 14-week AI skills program to their students in grades 11 and 12, covering foundational programming, AI and machine learning concepts, and real-world applications using industry-standard tools. Third, students will develop professional competencies, including communication, collaboration, leadership, and career readiness, through structured feedback activities, an industry speaker series, and a Make-a-Thon innovation challenge. Fourth, 20 students per year (60 total) will complete eight-week paid internships with partner technology companies and university research labs, translating classroom learning into direct workforce experience. A mentorship model pairing faculty, graduate students, and a Penn State Center for Science and the Schools (CSATS) faculty member with teachers and students ensures continuity across all levels of the pipeline. All curricular materials will be made freely available online, and findings will be disseminated through publications, conferences, and a public project website, ensuring the framework is accessible to educators and institutions nationwide. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-05

This project will test how the age of social partners in a population determines social behavior and adaptation. As individuals age, their own behavioral expression changes, meaning that populations with older members represent a substantially different environment than populations with younger ones. Population age structure can thereby determine how individuals interact with conspecifics, alter which traits determine fitness, and ultimately determine the strength and form of natural and sexual selection. Understanding how age structure generates and maintains behavioral diversity requires a multilevel approach that tests how individuals change their current phenotypic expression, as well as how selection shapes behavior over evolutionary time. This work will be the first to evaluate the importance of population age structure as a demographic driver of behavioral evolution. The project will advance NSF priorities in artificial intelligence by developing and using AI-assisted video tracking and morphological measurement and to quantify animal movement and social interactions at a scale that would be difficult to achieve by hand. These tools will allow the research team to extract fine-scale behavioral data from large numbers of beetles and build a more precise understanding of how social environments shape evolution. This project further develops a training pipeline to provide research opportunities in field biology for undergraduates, including community-college transfer students, expands K–12 evolution education, and strengthens public outreach in STEM. Using a well-established invertebrate model system, the forked fungus beetle, replicated laboratory and seminatural field experiments will manipulate population level demographic features to explicitly test which factors are most important in determining how individual behavior is expressed, how social networks form, and which traits are favored by natural and sexual selection. A series of behavioral assays will examine how social effects on behavioral expression scale up from individuals to social networks to population patterns of interaction. Dyadic interactions between individuals of different ages will be used to test whether age dependent behavior results from ontogenetic changes within focal subjects or plastic responses to partners of different ages. Focal subjects assayed in populations with manipulated age structures will test how behavioral plasticity results when individuals reside in group contexts with different age structures. A multi-year selection experiment will separately manipulate operational sex ratio, density, and age structure to quantitatively test which of these contexts has the strongest effects on individual and social selection targeting social behavior and social network phenotypes. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-04

How tetrapods, four footed animals, became terrestrial was a pivotal event in vertebrate evolution that set the stage for the diversification of tetrapods thereafter. The locomotor capabilities of early tetrapods are often modeled with extant salamanders since the latter have a generalized tetrapod body plan. Yet, salamanders exhibit tremendous morphological diversity across environments, providing a framework to assess the mechanical requirements for terrestrial locomotion by comparing morphological change across carefully matched evolutionary lineages. The greater effects of gravity may impose biomechanical constraints that preclude certain salamanders from moving on land, but the habitat that salamanders occupy differs between developmental strategies. Metamorphosis involves the development of an animal across two or more distinct life stages but can be biphasic (aquatic larvae to terrestrial adults) or multi-phasic (aquatic larvae to terrestrial juveniles to aquatic adults) whereas direct development remains in one environment. Thus, biomechanical constraints may be stronger in terrestrial direct developers than biphasic metamorphers since the former do not experience an aquatic stage. This project will integrate physiology, engineering, and evolutionary biology to examine how the interplay between habitat preference and developmental strategy affects the relationship between the structure and function of tissues (e.g., bones) and whole-organism performance (e.g., locomotion). Students will receive research training through a new Course-Based Undergraduate Research Experience on Organismal Form and Function and Professional Research Experience for Post-baccalaureates in Biology program. In addition, “Salamander Safaris” will be hosted during Amphibian Week to promote the participation of students in STEM. Locomotion places some of the highest physical demands (‘loads’) on bones and failure to withstand loads could cause fractures or even death in an animal, yet how bones evolved to support the loads imposed by aquatic vs. terrestrial environments is not well understood. Phylogenetic comparisons of whole-bone mechanics across ecologically diverse species will advance knowledge of how habitat and developmental strategy has shaped the evolutionary morphology of salamander limb bones. Investigators will quantify in vivo bone loading during terrestrial walking through synchronized 3D kinematics and kinetics. Investigators will then apply these loading data to collect the first dynamic measures of limb bone strength by integrating mechanical property testing and 3D digital image correlation. Finally, they will combine these techniques to examine how bone mechanics is affected by water-land and land-water transitions within a lifetime by comparing juveniles and adults from species with different developmental strategies (i.e., direct, biphasic, multiphasic). Bone strength is predicted to be highest in terrestrial direct developers lowest in paedomorphic aquatic salamanders, and intermediate in biphasic metamorphic salamanders. Stronger bones likely assist terrestrial species to withstand internal (muscle) and external (ground reaction forces) loads and then transfer this energy into propulsion. Compared to the femur, the humerus is expected to evolve at faster rates based on the multi-functional role of forelimbs (e.g., digging, reproduction, locomotion) that is likely less constrained compared to hindlimbs whose primary role is for generating propulsion. Findings from this work will contribute new insights into the mechanical requirements of becoming terrestrial. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-03

Nitrate is an ion composed of nitrogen and oxygen that is used in agricultural and industrial processes. Nitrate accumulation in water poses a serious threat to ecosystems and human health. A promising solution uses electricity to convert nitrate into chemical products. The conversion involves catalytic reactions at the surface of an electrode. Current technologies for nitrate conversion are inefficient because multiple reactions occur at once and processes at the electrode surface are not well understood. This project will find new ways to study reactions at the electrode surface and improve nitrate conversion. The results will help new developments for water treatment and chemical manufacturing. The project will engage high-school and undergraduate students in hands-on research and design-based learning that will stimulate interest in STEM careers. Electrocatalytic nitrate reduction and competing hydrogen evolution collectively involve a complex reaction network that is strongly influenced by catalyst structure and the surrounding solution environment. This project will combine quantitative kinetic analysis with time-resolved operando infrared spectroscopy to identify key surface intermediates and solvent structures during catalysis. Kinetic models that account for non-ideal electrolyte behavior will be used to quantify how electrolyte composition modifies the energetics of kinetically relevant elementary steps. These insights into the mechanisms of nitrate reduction and hydrogen evolution will enable designing new electrolyte formulations and catalysts that improve efficiency and selectivity to desired products. The project will establish broadly applicable strategies for linking interfacial structure, solvation, and reactivity in electrocatalysis, with relevance to a wide range of electrochemical reactions beyond nitrate conversion. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-01

NON TECHNICAL: Finding out how a particular processing path either purposefully creates or minimizes various classes of defects is key to further progress in exploiting soft matter crystals for their intriguing material properties. Soft matter crystals such as those comprised of block copolymer molecules form via self assembly of the component molecules into periodic arrays. Recent advances made with new types of microscopy now enable 3D visualization of the fine-scale features of organized block polymers. While the structures are mostly regularly periodic, defects occur and the aim is to use microscopy tools to see these local disruptions in the order, classify them by their geometry, learn how they form during processing of the material and how each type of defect influences material properties. The performance of periodic materials depends to a great extent on having well ordered structures and when desired, precisely positioned and aligned defects. The PI seeks to discover valuable types of defects that can then be manipulated to enhance and even create brand new technological applications of block copolymer materials. For example, a defect that forms a type of "molecular mirror" inside a structure without mirror symmetry may create new opportunities for manipulating the propagation of light waves. Since defects are relatively rare events, the search to discover and classify heretofore mostly unknown objects requires high quality, large data sets and use of machine learning for statistically sound and unbiased microstructural data analysis. Locating, analyzing and classifying all the various types of defects created under a given set of processing conditions will benefit materials researchers in fields well beyond polymeric materials. Such processing – structure relations have been the goal of materials science since its inception and while much has been learned about how different polymer processing procedures influence microstructure, the level of information has generally been limited to observations on polymer chain, domain and crystal orientation, domain shapes and periodicities has but rarely been related to the type, amount and distribution of the many types of defects throughout the material. TECHNICAL: The PI's group will utilize the advances made with slice and view dual ion and electron beam microscopy for reconstruction of very large volumes of 3D tubular network microdomain block copolymer samples to identify and characterize the defects and relate them to the processing route used to prepare the sample and to their influence on properties such as charge, mass and wave transport. Since defects are relatively rare events, the search to discover and classify heretofore mostly unknown objects requires near distortion-free microscopy to provide large data sets. The inherent complex topology and morphology of defects in network phases requires the help of machine learning for morphological analysis. Machine learning on these large data sets will afford statistically sound and unbiased microstructural data analysis for location and characterization of all the various defects with the ability to find the most abundant and also distinctive patterns of defects. Understanding how various types of defects are created under a given set of processing conditions will benefit materials researchers in fields well beyond polymeric materials. Development of improved processing protocols such as membrane homogenization of microparticles to avoid anisotropic sample deformations and perhaps even growing the first faceted true single crystals of block copolymers may demonstrate that polymers can realize near-perfection as do other classes of soft matter. Such processing – structure relations have been the goal of materials science since its inception and while much has been learned about how different polymer processing procedures influence microstructure, the level of information has generally been limited to observations on polymer chain, domain and crystal orientation, domain shapes and average periodicities has but rarely been related to the type, amount and distribution of the many types of precise defects throughout the material. Ultimately the research will discover valuable defects that can be manipulated to enhance and create new technological applications of block copolymers. . This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2026 · 2026-01

Artificial intelligence (AI) is becoming essential to scientific discovery in areas, such as biomedical research, environmental modeling, and genomics. However, the reliability of AI systems depends on the quality and integrity of the data used to train them. Scientific datasets are often collected from multiple sources, including laboratory instruments, simulations, and collaborative institutions. This variability makes it difficult to verify how data were generated, processed, or applied. This project supports the NSF's mission to advance trustworthy computing by developing an infrastructure that tracks the full lifecycle of scientific datasets using data provenance methods. By enabling end-to-end traceability, the work improves transparency and accountability in AI-driven science. The project introduces a three-layer architecture for capturing data provenance from hardware devices, operating systems, and scientific applications. The resulting provenance events are merged into a unified provenance graph that supports scalable storage and analysis across institutional boundaries. The research also develops privacy-preserving anomaly detection techniques using federated machine learning, allowing institutions to identify suspicious data behaviors without sharing sensitive raw data. To reduce barriers to adoption, the system includes an investigation interface that supports natural language queries over provenance graphs, helping researchers understand how data were used or potentially manipulated. Expected outcomes include open-source tools, curriculum materials for AI data integrity, and evaluation on real-world datasets from biomedical and environmental workflows. These results will enhance secure collaboration, foster reproducible science, and improve public trust in data-intensive research across academia, healthcare, and government. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-11

The project develops a design methodology for supported single-atom catalysts (SACs) – an emerging class of supported single metal-atom catalysts that offer exciting and emergent properties that can revolutionize many industrial applications. The realization of their full potential is hindered by limited understanding of how to control their stability and catalytic properties within the complex material design space extending across the properties of the metal atoms and supporting material, together with interactions between the two. To overcome this challenge, the project embraces a highly-integrated, computational-experimental methodology using machine learning techniques (ML) to leverage the support material as a ligand to regulate the geometric and electronic properties of the metal site and improve its stability. The model predictions will guide the synthesis, characterization and catalytic measurements to enable selective bond activation. The proposed methodology can profoundly impact the discovery of complex materials for challenging chemical reactions. The design of stable, active, and selective catalysts, while maximizing the metal utilization at the single-atom level, can significantly reduce capital costs and energy consumption, leading to lower CO2 emissions, reduced production of harmful byproducts, and more responsible utilization of hydrocarbon feedstocks. The interdisciplinary nature of this research and the integration of research and education plans between the three institutions will lead to a cadre of students obtaining a unique educational experience in heterogeneous catalysis, multiscale modeling, and advanced lab- and synchrotron-based characterization techniques. Furthermore, the project will develop educational materials for outreach programs targeting K-12 students with focused efforts to increase the participation of underrepresented students in STEM fields. The project incorporates a conceptual framework centered on artificial intelligence (AI) and multiscale modeling-based methodologies to build guiding principles that can be leveraged to predict highly active, stable, and selective metal-support compositions. The model predictions will guide the synthesis of single-metal atoms supported on novel, high-surface-area unconventional support materials (perovskites and spinels) by atomic layer deposition, followed by detailed characterization of their properties, catalyst evaluation, and model assessment and refinement (thus enabling an efficient catalyst discovery/design loop). By uncovering physics-inspired descriptors and harnessing the capabilities of machine learning, the project aims to predict how the surface composition of the oxide support and the local cation environment at the metal site influence stability, activity, and selectivity. The developed methods and models will be evaluated with respect to two complex industrially relevant reactions: 1) water-gas shift, and 2) hydrodeoxygenation (HDO) of cresol to toluene. The former focuses primarily on maximizing reaction rate, while the latter addresses both activity and selectivity challenges. The outcome of this research will serve as a foundational methodology for designing new materials in silico. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-11

Catalysts are used widely in industrial processes to improve energy efficiency and direct chemical reactions toward desired products. The project examines novel catalyst designs for an important class of industrial reactions – selective hydrogenations - used for many applications such as the production of plastics, fragrances and pharmaceuticals. The precious metal palladium (Pd) is widely used for selective hydrogenation reactions. The project aims to maximize utilization of the precious metal in the form of isolated atoms, and tailor their properties by anchoring them to covalent organic frameworks (COFs), which are highly crystalline porous materials with specific binding sites for the metal atoms. The investigators will synthesize and characterize the catalysts using a suite of laboratory- and synchrotron-based characterization techniques. The catalysts will be tested for the selective hydrogenation of acetylene to ethylene to determine how the binding sites affect the activity and selectivity of the Pd atoms. Beyond the research, the project involves education of graduate and undergraduate students with cross-training of researchers between the two institutions and focused efforts to increase participation of underrepresented groups in catalysis science. The project aims to tailor the properties of Pd single atoms using covalent organic frameworks (COF) as the support/ligand, and thus circumvent trade-offs between activity and selectivity associated with traditional metal oxide supported nanoparticle catalysts for the semi-hydrogenation of ethylene. Areas of focus include 1) the effects of the COF binding sites and secondary ligands on the Pd electronic properties, 2) hydrogen activation, 3) binding with adsorbates and 4) catalyst activity/selectivity. Those thrusts will be accomplished through precise catalyst synthesis, advanced characterization techniques (in-situ/operando X-ray spectroscopies, microcalorimetry) and detailed kinetic measurements. The project will identify the possibilities and limitations for designing catalysts based on Pd metal single atoms. The interdisciplinary nature of this research, and the integration of research and education plans will lead to a cadre of students obtaining a unique educational experience on heterogeneous catalysis and advanced lab- and synchrotron-based characterization techniques. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

Non-technical Abstract: The magnetic properties of materials originate from the magnetic atoms that compose them. These atoms act like tiny bar magnets. In materials such as iron, the atomic magnets align parallel to one another, much like compass needles pointing in the same direction. However, in some materials, the atomic magnets arrange themselves in non-parallel, or non-collinear, configurations. These arrangements arise from the interplay between the crystal structure and intrinsic interactions among the magnetic atoms. Certain non-collinear magnetic structures possess topological properties—unique geometric characteristics that make them robust against deformation and disruption. This topological protection holds significant promise for applications in information storage and quantum computing. This project combines experimental and theoretical approaches to study topologically protected magnetic structures, aiming to identify materials that could serve as platforms for future quantum technologies. Our research focuses on a remarkable class of compounds known as Heusler alloys—chemical combinations of several metals, including magnetic elements like iron. Recent discoveries have shown that Heusler alloys with specific crystal structures can support non-collinear magnetic arrangements. Due to their relative ease of synthesis and tunable magnetic properties, Heusler alloys offer a conducive environment for the discovery of topologically protected magnetic phases. In addition to advancing quantum science, this project will provide undergraduate and graduate students with vital experience in cutting-edge quantum research, helping train the next generation of quantum scientists and engineers. Technical Abstract: The main goal of this project is to identify the relationship between crystal structure, chemical composition, electronic band structure, and topologically protected magnetic states to design / discover novel quantum materials from the Heusler family of alloys. These materials are actively studied for practical applications such as spintronics, quantum information science and engineering, data storage, magnetic cooling, shape memory and magnetocaloric devices. Exploring topologically protected magnetic phases, such as skyrmions and antiskyrmions, as well as other forms of magnetic non-collinearity in Heusler compounds for obtaining fundamental understanding of these phenomena, which can then be applied to the development of practical device applications including novel data storage mechanisms, constitutes the main research objective of this project. The main hypothesis of this project is that Heusler materials with tetragonal crystal structure may exhibit non-collinear magnetic order, which may in certain cases result in topologically protected magnetic phases, such as skyrmions and antiskyrmions. The research team is using various experimental and theoretical techniques to perform the project, such as arc-melting, physical vapor deposition, Lorentz transmission electron microscopy (LTEM), electron-transport measurements, and density functional theory (DFT) calculations. The project aims to uncover the underlying physical principles of topological magnetic states and other forms of magnetic non-collinearity in Heusler materials. This allows the research team to identify / discover mechanisms to control these properties by intrinsic chemistry change or other forms of external stimuli (such as mechanical strain) leading to the discovery of new quantum materials exhibiting such properties. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

The rotation of the planet, the stratification of density, and the generation of friction all play important roles in the motion of geophysical fluids. These mechanisms are always present in our ocean and atmosphere, as well as that of other planets. They are collectively responsible for many well-known phenomena that we observe in nature, e.g., jet streams, zonal jets, the El Niño cycle, and Jupiter's Great Red Spot, to name only a few. These mechanisms typically serve to constrain the motion of the fluid in a very particular way. For instance, it is observed that in a rapidly rotating fluid in three-dimensions, particles that are aligned along a common vertical parallel to the axis of rotation move nearly in unison, thus rendering the overall motion of the flow to be essentially two-dimensional. Despite many experimental and computational efforts to understand the precise development of such phenomena, the mathematical justification for them, that is, from directly studying the equations of motion themselves, remains largely open. This project will systematically address such concerns in various geophysical settings. This project will also provide research and mentorship opportunities for students at the undergraduate and graduate levels, as well as postdoctoral scholars. An overarching goal of this project is to understand various manifestations of finite-dimensionality and its interconnections with the mechanisms of dissipation, rotation, and stratification. The main approach will be through the study of the regularity and long-time behavior of solutions to the associated equations of motion that allow one to obtain precise quantitative relations between the parameters representing the strength of these various mechanisms with the smallest relevant length scales of the fluid flow. The main models of interest will be those that arise naturally in geophysics such as the rotating Navier-Stokes equations and the stably stratified Boussinesq equations. In order to properly quantify the effects carried by rotation and stratification, anisotropic dispersive estimates, and careful analyses of resonance structures inherent in such systems will be carried out. A novelty of this project is the interplay between physical space- and frequency space-based approaches. Although both approaches have seen success in studying the regularity of solutions, the frequency space-based approach is well-suited for quantifying the number of degrees of freedom, while the physical space-based approach is well-suited for exploiting information about the spatial analyticity radius. This project attempts to merge these two approaches in ways that allow one to jointly exploit the reductions in dimensionality in both physical-space and frequency-space that is observed in rotating or stratified fluid flows. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

Non-technical Description: Energy storage technologies such as batteries critically require safe and thermally stable ion-conducting materials. Lithium-ion batteries are pervasive in modern society, but safety concerns have prompted the development of new solid-state ion-conducting materials. To date, nearly all solid polymer ion conductors are comprised of synthetic materials that lack precisely defined structures. In contrast, biological macromolecules such as peptides have precisely defined sequences, allowing for control over three-dimensional molecular structure, such as helical elements. This project aims to understand the role of peptide helices on enhanced ion conduction, focusing on: (1) the ability to arrange ion-conducting groups in controlled ways and (2) the presence of a macrodipole along the backbone that grows with helix length. This project will enable the development of a new class of helical peptide ion conductors for enhanced energy storage applications. A wide range of peptide chemistries will be designed and synthesized using an AI-guided discovery approach to understand the role of chemistry, sequence, helical character, and arrangement of ion-conducting groups on the performance of energy storage materials. A key outcome of this project is to understand how the molecular structure of peptides affects ion transport for the development of next-generation energy storage materials. Technical Description: This project will address key challenges in developing new materials for enhanced ion conduction by efficiently exploring a vast chemical space using a machine learning (ML)-guided approach, together with complementary materials synthesis and characterization methods. The team focuses on developing a new class of helical peptide-based polymers for enhanced ion conduction by controlling the macrodipole inherent to peptide-based helices and the spatial arrangement of ion-conducting motifs away from the backbone. To this end, this project will leverage the ability to control several key properties of peptide electrolytes such as: (1) helicity via introduction of amino acids of opposite chirality; (2) molecular weight via controlled ring-opening polymerizations of cyclic amino-acid monomers; (3) ion conducting motifs and linkages to polymer backbones using non-natural amino acids; (4) monomer sequence via solid-phase peptide synthesis; and (5) alignment of helices via materials processing, e.g., hot-pressing at different temperatures and pressures. PIs and graduate students will engage in outreach activities geared towards hands-on demonstrations of scientific concepts for middle and high school students. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

This study examines the impact of non-science isolation practices on scientific production and collaboration. The research deepens our understanding of how national scientific systems are interconnected with global economic and political dynamics, providing insights on the impact of scientific cooperation across countries and sectors. The study addresses fundamental issues at the intersection of science strategy, international relations, and research productivity including how restrictive measures and geopolitical shifts affect international scientific collaboration, researcher mobility, and resource exchange. The project serves the national interest by promoting the progress of science through a better understanding of the factors that influence global scientific cooperation. The results provide decision makers with empirical evidence to inform strategic decision-making in the governance of scientific activities. This knowledge is crucial for maintaining the United States as a leader in global scientific research and innovation. The findings from this study have broader impacts on national health, prosperity, and welfare by helping to navigate international scientific collaboration in an increasingly complex geopolitical landscape. The researchers employ quasi-experimental designs utilizing three recent global events. These events serve as cases to examine the causal impact of disengagement from the integrated global scientific community on scientific development. The study will use causal inference techniques, specifically the synthetic control method and difference-in-differences analysis. These methods are applied to investigate changes in scientific publication production and the exchange of scientific capital, including funding resources and researcher mobility. To address potential biases in international databases, the project constructs a comprehensive dataset incorporating publication records from both international and national sources. By using causal inference frameworks, the researchers offer nuanced insights into the consequences of withdrawing from the integrated scientific community while mitigating the influence of other factors. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

Recently, the rapid emergence of quantum optics for quantum communications, signal processing, quantum sensing, and next-generation computing has revealed tantalizing images of imminent technological breakthroughs. However, many of the requisite components do not exhibit the necessary levels of performance, or practical implementations do not yet exist. This is particularly true for one of the key components, single-photon detectors, where the high detection efficiencies (approaching 100%) required for these applications have not been achieved. To date, the most widely used single-photon detector types are superconducting nanowire single-photon detectors (SNSPD) and single-photon avalanche diodes (SPADs). SNSPDs have demonstrated excellent performance but must be operated at < 4 K. SPADs have the advantage of working near or at room temperature. Additionally, SPADs are compact, low-cost, and suitable for deployment in arrays, operating over a wide spectral range from ultraviolet to infrared. Due to fundamental material limitations, state-of-the-art SPADs have not achieved single-photon detection efficiencies exceeding 60%. This program will use a new SPAD material, AlInAsSb. It possesses a unique capability to achieve an ultra-high avalanche probability. Combined with a waveguide structure for high collection efficiency, the high detection efficiencies required for quantum optics applications (~ 99%) can be achieved at room temperature. Previously, SPADs have not achieved photon number resolution, a crucial requirement for quantum computing. This program will develop a novel segmented detector consisting of a linear array of nano-SPADs that can achieve number resolution and high efficiency. Technical Description: A primary thrust of this program is to achieve single-photon avalanche diodes (SPADs) with photon detection efficiencies > 90%, ultimately approaching 99%. We will use a new SPAD material, AlInAsSb. The photon detection efficiency is the product of the external quantum efficiency and the avalanche breakdown probability. For normal incidence devices, our AlInAsSb APDs already have the same external quantum efficiencies as the InP/InGaAs APDs that have been widely used as SPADs at 1550 nm. However, to achieve ~ 99% external quantum efficiency, we will develop waveguide structures that enable high efficiencies since they have longer absorption lengths. We have already demonstrated PIN waveguide detectors with an efficiency exceeding 95%. The advantage of AlInAsSb is that its disparate electron and hole impact ionization coefficients enable high avalanche breakdown probabilities; greater than 90% has already been achieved. The state-of-the-art InP/InGaAs SPADs are limited to ~ 65% by their ionization probabilities. In summary, our approach with AlInAsSb is the only means to achieve the ultra-high detection efficiencies required for quantum optics applications. Previously, SPADs have not achieved photon number resolution, a crucial requirement for quantum computing. This program will develop a novel segmented detector consisting of a linear array of nano-SPADs that can achieve number resolution and high detection efficiency. Another key parameter for SPADs is the dark count rate. Current AlInAsSb SPADs exhibit dark current densities approximately 10 times higher than InP/InGaAs SPADs. A variety of approaches, involving three groups of collaborators, is being pursued to reduce the dark current. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

This joint NSF-ANR (France) project will tackle theoretical and experimental studies of multipartite entanglement of optical beams with the aim of demonstrating novel technologies in quantum sensing and quantum communication. Multipartite entanglement can be viewed as “3-body” (or more) quantum correlations and is the key feature of disruptive quantum technologies, first among them quantum computing. On the sensing side, GANDALF aims at generalizing, to multiple beams of light, the measurement sensitivity increase achieved for a single beam of light by the Laser Interferometer Gravitational-wave Observatory (LIGO). On the communication side, GANDALF will enable classically unfeasible multipartite quantum communication protocols such as quantum secret sharing, which encodes secret information in a way that can only be unscrambled by the honest collaboration of all communication partners. The GANDALF team proposes to demonstrate multipartite entanglement both in free-space optics (US) and integrated optics (France) and apply them to quantum sensing and communication protocols which are not realizable by classical physics. These experimental efforts will have different foci: bright beams in on-chip photonics in France, non-Gaussian faint states of light for quantum error correction in the US. Moreover, GANDALF’s theory team will pursue fundamental studies of continuous-variable (i.e., field-encoded) quantum information which will inform both experimental efforts. The ultimate goals of the project are to scale up the number of multipartite entangled beams and to progress toward technology deployment in the field by implementing quantum error correction. This collaborative U.S.- U.K. project is supported by the U.S. National Science Foundation (NSF) and the Engineering and Physical Sciences Research Council (EPSRC) of United Kingdom Research and Innovation (UKRI), where NSF funds the U.S. investigator and EPSRC funds the partners in the U.K. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

This project advances national health and promotes science and technology development by providing algorithms, software, and systems that can train machine learning models on electronic health records (EHRs) for accurate and early prediction of Alzheimer’s Disease and Related Dementias (ADRD). ADRD is a severe neurodegenerative disorder that effects over 5,000,000 people over the age of 65 that is characterized by progressive memory, cognitive impairment and personality changes, which can further evolve to dementia and death. Early prediction of ADRD is crucial for timely intervention and improved patient outcomes. Recent studies have shown that personal risk factors such as education, employment, and lifestyle or family history significantly influence ADRD onset and progression. However, these factors are not recorded in a structured format within the existing EHRs. In contrast, personal risk factors are often embedded within the free text of clinical notes or discharge summaries that are not easily searchable, computable, or standardized. This creates a major technical barrier for their integration into the ADRD prediction models. To address this, this project develops a computational platform using novel machine learning and natural language processing to automatically extract personal risk factors from EHR clinical narratives and leverage them for accurate and early prediction of ADRD. This research significantly improves ADRD prediction accuracy and timeliness, with potential generalizations to other neurological disorders. By exploring the interaction between personal and clinical factors in disease development, this project pushes the boundaries of current knowledge in machine learning and ADRD research, potentially transforming approaches to early detection and management of complex neurological disorders. To achieve the goal of developing personal risk factor enhanced machine learning models for early ADRD prediction, this project develops four thrusts of novel approaches, each addressing key methodological challenges. First, the project develops a domain knowledge guided large language model to extract risk factors from EHR clinical narratives, which can adeptly cope with the complexities inherent in real world EHR clinical narratives, such as noise and incomplete data entries. Second, the project develops an interpretable method using neural additive models that automatically identifies the individual risk factor’s contribution to the early ADRD prediction. Building upon this interpretable result, in the third thrust, the project develops a survival-based ADRD prognosis model that can be used to estimate the likelihood of ADRD development at any given point in the future, capturing the dynamics of risk trajectory. This approach can enhance clinical decision-making by identifying high-risk individuals who may benefit from more intensive care or early intervention. Fourth, this project constructs a personalized knowledge graph that integrates personal and other clinical risk factors into a unified format for capturing the overall health status for everyone at risk of developing ADRD. Moreover, this project develops adaptive machine learning algorithms that can dynamically update this knowledge graph to incorporate the evolving risk factors. Together, these approaches converge to address the fundamental limitations of existing ADRD risk prediction models, such as inability to handle complex and unstructured data, insufficient interpretability, and high computational overhead. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

This Level 3 IUSE project, a collaborative effort between Florida International University (FIU), California State University San Marcos (CSUSM), Clark College (CC), and the University of Virginia (UVA), aims to improve teaching and learning in introductory calculus courses. Specifically, the project adapts and implements the Modeling Practices in Calculus (MPC) approach, which has been shown to significantly improve learning and success for a broad collection of Calculus I students at FIU. At the core of MPC is the opportunity for students to work cooperatively and discuss key mathematical ideas in a face-to-face, small group setting. The curriculum engages students in the practices of mathematicians, supports students argumentation, and builds important mathematical reasoning and communication skills. This project implements MPC at multiple sites across two overlapping phases, first facilitating adoption and implementation at CSUSM, CC, and UVA, and then identifying three additional partner sites that will join the project in its second year. Calculus remains a key course for many STEM degree programs and can represent a bottleneck for many students. This project uses evidence-based teaching and engagement practices to provide an enriching, collaborative learning environment that has potential to improve student learning and outcomes, thereby increasing the number of students successfully completing STEM degrees in multiple areas of national importance. Project research focuses on three research questions: 1) Faculty focus: How do faculty develop their MPC instructional expertise and what professional development elements are necessary for future propagation efforts? 2) Student focus: How do students develop cognitive and affective skills arising from the adapted MPC instructional designs? 3) Propagation focus: How are Phase 1 partners prepared for establishing a national calculus propagation network? How are new sites successfully recruited and onboarded into project? Data sources include interviews, surveys, weekly faculty reflections, and shared exam questions across project and non-project courses. Project evaluation, carried out by an experienced evaluator and supported by an expert advisory board, will assess project implementation, project and site operations, and overall progress towards goals. Outcomes and findings will be disseminated broadly to encourage further adoption of the MPC approach across institutional contexts. The NSF IUSE: EDU Program supports research and development projects to improve the effectiveness of STEM education for all students. Through the Engaged Student Learning track, the program supports the creation, exploration, and implementation of promising practices and tools. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

Ensuring modern computing systems are secure is critical for preventing malicious adversaries from intercepting data, compromising infrastructure, and stealing intellectual property. Today’s computers rely on dedicated security chips as the foundation for overall system-level security. Increasingly, vendors are using the Tock operating system on these chips because Tock solves a key need in this industry: secure and reliable software. The security of computing devices (e.g., datacenters, laptops, and smartphones) then ultimately relies on Tock’s correctness and robustness, and Tock’s development ecosystem must remain trustworthy and resistant to compromise. This project will develop new techniques to ensure that malicious attackers cannot introduce covert flaws into the Tock project that would undermine the robustness and security of the operating system. This project will provide layered defenses against increasingly patient and capable adversaries using techniques including automated robustness analysis of software dependencies; compatibility checkers between hardware, the operating system, and applications; and open-source contribution review aids. In addition to ensuring Tock remains a trusted platform for security chips, these advancements could be generalized to support the robustness of other high-assurance software projects. Vendors rely on Tock because it is implemented in a memory-safe language (Rust), provides robust isolation and least-privilege guarantees, and is open-source. However, as an embedded operating system, Tock must support and directly interface with numerous hardware platforms while meeting resource constraints. These requirements mean that existing system-, software-, and language-level features for reliability and correctness are inadequate, and a large contributor base is needed to support diverse hardware platforms. This makes Tock susceptible to covert development attacks targeting the software supply chain, the trusted-computing-base, and the open-source code review practices for low-level systems code. This project will develop mechanisms to protect the Tock open-source project. These include detectors for unsafe and unsound code in Tock and its dependencies, automatic generation of system call interfaces that prevents exploitable incompatibilities between user space and the kernel, and code trust tiers to guide manual code review scrutiny to critical components. These defenses will enhance the guarantees of using a memory-safe language for low-level code while addressing weaknesses in the open-source contribution model. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-10

The success of deep neural networks (DNNs) and large foundation models such as vision language models (VLMs) and large language models (LLMs) has called for incorporating them in sensitive domains, such as medical applications. However, widespread clinical adoption remains extremely limited. Prototypes of these systems suffer from fundamental problems, such as the black-box nature of AI models; that is, a lack of transparency and interpretability. One critical issue that holds back the widespread adoption of machine learning in healthcare is the lack of accountability in the system's predictions. To make these models useful for healthcare, these systems must be designed with a bedrock of solid accountability and sound rational decisions before focusing on raw performance metrics. This EArly-concept Grant for Exploratory Research (EAGER) project develops concept-based reasoning approaches to improving interpretable DNN model design and improving rationales in large pre-trained models. The concept-based reasoning attempts to learn high-level ‘concepts’, which are abstract entities that align with human understanding and provide explicit rationales for model predictions to achieve interpretability. This EAGER project conducts two innovative research tasks: (1) Build inherently explainable concept-based DNN Models for Health. Concept learning models attempt to learn high-level concepts, abstract entities that align with human understanding, and thus provide interpretability to DNN architectures. This approach can be effectively utilized in domains such as medical diagnosis where concepts are sometimes undefined, for example, shape of bone in an X-ray where the shape itself is undefined, but a doctor can easily tell it is irregular. (2) Improve the pre-trained foundation models’ trustworthiness by incorporating concepts as structured entities that provide human-understandable reasoning. This task designs a neurosymbolic framework which provides concept-based reasoning through learning visually grounded concepts, linked neurosymbolically through rules. The framework is particularly effective on medical data, which is typically underrepresented in the pre-training of the large VLMs. A hierarchical concept tree structure is designed to provide a symbolic reasoning process through a vast search space created using knowledge banks to improve VLM predictions during inference. This research has the potential to have a significant impact in designing state-of-the-art medical diagnostic systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-09

Artificial intelligence - in particular, deep learning - is rapidly being developed for healthcare systems with great potential to improve disease diagnosis, treatment planning, and patient monitoring. However, the translation of these powerful models from research and development to everyday clinical use is being held back by the lack of trustworthiness in these systems. This project will explore strategies and develop methods for ensuring the robustness of deep-learning models in healthcare applications. A major obstacle to guaranteeing the behavior of deep-learning systems in healthcare is the wide variability in data across different healthcare sites, including a range of medical-imaging devices, data-collection protocols, and patient demographics. This can lead to data inputs to the system that are significantly different in nature from the data on which it was trained. To address this issue, we propose to develop robustness audits that assess how well a healthcare deep-learning system tailored to a specific site will operate at another site. Broader-impact aspects of the work include the potential to significantly and widely improve the effectiveness of deep learning in practical healthcare applications. Additionally, an array of educational and outreach activities are planned. The first goal of this project is to develop robustness audits using synthetic data that provide full coverage of test cases simulating conditions at a target healthcare site. This will be done by developing deep generative models with the ability to produce highly-realistic synthetic medical images that closely mimic the properties of imaging data collected at a target site. The second goal of this project is to use this generative-model framework to develop verification tools for measuring the robustness of a deep-learning healthcare system. This robustness will be expressed as regions in the latent space of the generative model, thereby restricting the set of data inputs to only valid medical images. The third goal of this project is to design natural-language models for communicating the results of the robustness audits to doctors. These models will produce textual descriptions of the conditions that would lead a deep-learning system to produce incorrect results at a particular site. Finally, the robustness-auditing system will be validated on real-world medical-imaging data in a cardiac-resynchronization therapy-planning task, where the deep-learning system to be audited predicts the optimal placement of pacemaker leads from magnetic-resonance imaging of the heart. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-09

One of the biggest gaps in our understanding of the earliest galaxy assembly, is the origin of supermassive black holes. Possible candidates include black holes that collapse directly from gas, as well as those that form from the earliest generation of stars, remnants of runaway stellar collisions, or black holes in dense clusters of stars. These channels are expected to take place in very early times. However, they are challenged by the masses and number of black holes that are observed at these early times. Researchers at the University of Virginia will implement novel theoretical models for the formation of supermassive black holes that will be crucial in interpreting the observations, with the goal of unveiling the SMBH origins. As part of the project, the researchers will build a teacher training program to develop related classroom content for K-12 teachers. Current cosmological simulations cannot fully leverage the upcoming wealth of gravitational wave and electromagnetic observations. A key limitation of large cosmological simulations is their simplistic treatment of the multiphase interstellar medium (ISM) via an effective equation of state. The key science goal is to build a multi-scale suite of cosmological hydrodynamic simulations that will identify distinct gravitational wave and electromagnetic signatures of Population III, nuclear star cluster and direct collapse black hole seeds, detectable within the James Webb Space Telescope and next generation facilities like ngVLA, extremely large telescopes and LISA. To achieve this, a large systematic study of the formation and growth rates of different seeds within explicit-ISM simulations will be conducted, accompanied by a characterization of the host galaxies of seed descendants and comparisons with recent observations. The work should lead to predictions of the LISA gravitational wave event rates for the various black hole seeding channels. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-09

In the “Beyond Moore’s Law” era with increasing edge intelligence, domain-specific computers in heterogeneous fabrics will rule the roost. Algorithms accelerating NP-hard (i.e., provably complex) applications or pre-compute processes that do not demand exact precision will run on tailored hardware. The hardware performance, rather than the algorithmic or software efficiency, may dictate solution speed, energy cost, footprint, and cyber-resilience. Clever hardware innovations for application-specific integrated circuits (ASICs) are no longer a rarity, but they all employ conventional material platforms like silicon, insulators, and compound semiconductors. This proposal will explore a new prospect – the use of quantum materials with exotic properties – to elicit computational activity with unprecedented size, weight, and power (SWaP). Additionally, innovative technologies and methods to train students in lab procedures through virtual platforms (e.g. GoPro video sessions, kid-friendly Minecraft and Roblox design challenges) will be developed and posted on YouTube and Vimeo for the public. Students selected through online exercises will be evaluated using rubrics developed by learning centers at the universities and sent to the Army Research Laboratory (ARL) and the National Institute of Standards and Technology (NIST). For the hardware needs of modern computing and artificial intelligence to be “self-contained”, all the data and resources needed to execute a computing task should be available in situ and not have to be fetched from a remote server or “cloud” which may be unreliable or unavailable. One powerful paradigm that satisfies many or all of these requirements is “processor-in-memory (PiM)”, where compute happens right at the memory site. The project plans to design, simulate, fabricate, characterize, and experimentally demonstrate a processor-in-memory architecture implemented by heterogeneously integrating a topological insulator (TI) (a quantum material) with nanomagnets and a piezoelectric material. The nanomagnet enables storage and the piezoelectric enables gating, while the TI brings in both high spin-selectivity and voltage tunable bandgap. The device is projected to perform logic operations and image processing with ultralow footprint and energy cost. A new and powerful PiM-based on a novel genre of materials with unusual quantum mechanical properties will be developed, which can be leveraged to outperform other PiMs in energy consumption, footprint and speed. This PiM will be built, characterized and its superior performance demonstrated. New light will be shed on the physical properties of these quantum materials to stimulate further research to benefit computing hardware. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.