University Of California Riverside

NSF Awards · FY 2026 · 2026-11

A Programming Languages Mentoring Workshop (PLMW) is organized as part of the ACM SIGPLAN Symposium on Principles of Programming Languages (POPL), the flagship conference in the field of programming language theory, and one of the premier conferences in all of computer science. The 2027 conference will be held in Mexico City, Mexico. Many POPL papers directly address administration priorities in artificial intelligence (AI) and Quantum computing. The impact of the award relates to providing opportunities for students to receive mentoring from leading researchers, and building the next generation of researchers and knowledgeable practitioners in programming languages. The award's broader significance and importance include building international community, lasting professional connections to design novel programming languages and implement tools, and enhancing education of US students. The workshop also provides students exposure to and multiple opportunities to interact with leading-edge research and researchers. By supporting students, the workshop thus imparts training to the next generation of researchers in programming languages and systems and contributes to building a national workforce in these topics. 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

Dry Eye Disease affects over thirty million Americans and imposes an annual economic burden exceeding $55 billion in healthcare costs and lost productivity. Despite its clinical importance, the fundamental principles connecting the molecular composition of this lipid layer to its mechanical performance remain poorly understood. Most current treatments address symptoms rather than the underlying causes of film instability. This research project will identify the design rules that govern how tear film lipid composition controls stability, lubrication, and resistance to mechanical failure. The findings will provide a scientific foundation for developing targeted therapies for Dry Eye Disease. They will also guide the creation of bio-inspired lubricating materials for medical technologies such as long-wear contact lenses, ocular implants, and artificial joint systems, contributing to United States leadership in biotechnology. The project includes a hands-on outreach module, The Amazing Engineering of Your Eye, designed to introduce kindergarten through K12 students to core concepts in biomechanics and biomedical engineering. University students will receive training in advanced biophysical measurement methods, building a skilled and interdisciplinary workforce for the biotechnology sector. This research project seeks to establish a quantitative, mechanism-based framework linking the molecular architecture of the tear film lipid layer to its interfacial biomechanical and tribological performance. Although the biochemical composition of the tear film has been extensively cataloged, the causal relationships between nonpolar lipid attributes and the resulting mechanical stability, adhesion, and frictional response remain unresolved. Three coordinated research objectives address this gap through the lens of interfacial biomechanics and mechanobiology. Objective 1 will determine how structural features of major nonpolar lipid classes, wax esters, cholesterol esters, and triacylglycerols, regulate thermodynamic stability and interfacial viscoelasticity using Langmuir trough isotherms and dilatational rheology. Objective 2 will map the molecular-scale adhesive energy landscape between the lipid layer and the lid-margin glycocalyx using the Surface Forces Apparatus, enabling simultaneous measurement of piconewton-scale normal forces and sub-nanometer separations via multiple-beam interferometry. Objective 3 will identify the dominant tribological mechanisms responsible for ultra-low friction under high-speed cyclic shear and test the distinct biomechanical roles of lipid classes. These studies will produce the quantitative force maps for this biologically critical interface and establish predictive structure–function relationships connecting lipid molecular architecture to interfacial mechanics, adhesion, and lubrication. The findings will enable a new mechanistic foundation for understanding biomechanical failure of biological thin films, with implications for ocular surface mechanobiology, tribology, and the rational engineering of bio-inspired lubricating interfaces. 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

Modern scientific, medical, and policy decisions increasingly rely on observational data to evaluate the effects of interventions such as medical treatments, public programs, and behavioral exposures. Many of these interventions are multi-level or continuous, such as medication dosage or program participation intensity, rather than simple binary choices. However, existing statistical methods, including widely used propensity score approaches, often become unstable or unreliable in these settings, particularly when applied to high-dimensional data with complex relationships. This project addresses these challenges by developing new methods for reliable and interpretable causal analysis. The research will enhance the ability to draw credible conclusions from large-scale data, with applications in public health, healthcare delivery, social policy programs, and digital health. By improving the stability and accuracy of treatment-effect estimation, the project supports evidence-based decision making that can improve health outcomes, inform policy design, and promote societal well-being. In addition, the project will contribute to workforce development through student training, interdisciplinary collaboration, and outreach activities that broaden participation in data science and statistics. This research advances the theory and methodology of causal inference for complex data with multi-level and continuous treatments, where classical inverse propensity weighting methods can fail due to unstable weights and ill-conditioned estimation. It develops a unified and scalable framework for treatment-effect estimation beyond binary interventions by leveraging stabilized weighting, empirical likelihood, and deep learning representations. The project introduces new methodologies for counterfactual distribution estimation, dose–response analysis, and general loss-based causal inference. The proposed approach replaces unstable inverse weighting with stabilized weighting strategies, including empirical likelihood weighting and minimum-variance weighting, which control variance inflation and prevent extreme weights. It further incorporates modern deep architectures, such as Transformer-based models, to flexibly estimate nuisance functions in high-dimensional settings. Theoretical contributions include non-asymptotic generalization bounds, convergence rates, and asymptotic inference guarantees, establishing a rigorous foundation for combining deep learning with semiparametric causal inference. The project will also develop easy-to-implement software packages to facilitate practical implementation and broad dissemination of the proposed methods. The resulting methods will be validated on large-scale observational datasets, providing robust, interpretable, and scalable tools for causal learning in complex real-world environments. 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

Artificial intelligence (AI) is becoming integral to manufacturing, healthcare, and autonomous systems, creating an urgent need for reliable deployment across diverse computing platforms. Yet dependable deployment remains difficult because the software systems that adapt learned models to target devices are complex and fragile. An AI model that appears valid at a high level can still fail during deployment because of hidden resource limits, data layout requirements, and platform-specific transformations. These failures are especially concerning because they may silently alter outputs rather than cause visible crashes, making them difficult to detect, diagnose, and prevent. The project's novelties are new testing foundations that uncover hidden sources of deployment failure and assess whether AI systems produce consistent results across diverse hardware platforms. The project's broader significance and importance are improved reliability and trustworthiness of AI systems deployed in high-impact settings. The project also creates interdisciplinary educational opportunities through open tools, curriculum materials, and training that strengthen workforce development in dependable AI systems and heterogeneous computing. This project develops a cross-layer framework for testing the software stack that translates and executes deep learning models on heterogeneous hardware. The research has three integrated components. First, it develops methods for automatically mining parameterized constraints from model specifications, system implementations, and hardware resource limits, thereby exposing implicit assumptions that existing testing techniques do not capture. Second, it introduces targeted constraint negation as a new form of test guidance, driving test generation toward failure-prone regions while filtering out invalid inputs before expensive end-to-end runs. Third, it develops equivalent model rewriting and backend-aware differential checking to detect latent inconsistencies across platforms that should preserve the same model behavior. Together, these advances establish new correctness foundations for dependable AI deployment on heterogeneous computing 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 2026 · 2026-07

Fluoride ion batteries are a promising class of energy storage devices that use earth-abundant elements such as fluorine, calcium, magnesium, and copper. These materials are widely available and avoid reliance on critical minerals like lithium, cobalt, nickel, and vanadium, which have limited supply chains and are subject to geopolitical risks. The growing demand for reliable, cost-effective, and safe stationary storage is driven by rising electricity consumption, aging infrastructure, natural disasters, and the rapid expansion of artificial intelligence data centers. These facilities require uninterrupted power to manage variable loads and reduce downtime. This CAREER project addresses these challenges by advancing the scientific foundations of fluoride ion batteries and training a new generation of engineers and scientists. The researchers will investigate how fluoride ions will interact with the other battery components at a molecular level. The educational plan includes undergraduate research mentoring, curriculum development, and community outreach programs. Three new initiatives at University of California, Riverside (UCR) will be launched: National Lab Day to introduce students to research careers; the NanoScience Educator Workshop to train high school teachers; and a Day in the Lab experience to engage middle school students in hands-on STEM learning. The goal of this research is to understand how solvation structures and interphase formation influence fluoride ion transport and electrode reversibility in fluoride ion batteries. The project will integrate cryogenic transmission electron microscopy, advanced electrochemical methods, and synchrotron-based X-ray techniques. Three research objectives will be pursued. First, solvation structures will be designed using concentration tuning and hydrogen bonding among fluoride salts, solvents, and polymer matrices to promote stable interfaces. Second, the project will explore fluoride salts containing large organic cations to improve solubility, ion conductivity, and interphase properties in low-polarity solvent systems. Third, the interphase formation and failure mechanisms of electrode materials based on earth-abundant elements will be examined to improve electrochemical reversibility. This research will generate new knowledge in electrolyte and interface chemistry and help establish fluoride ion batteries as a viable platform for stationary energy storage in future grid and data center applications. 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

This project will create a safe and reliable way for people and robots to work together through teleoperation (remote control of machines), especially in high-risk settings such as surgery, disaster response, and space or nuclear operations. In these environments, robots must perform precise tasks while responding to human guidance. However, many current artificial intelligence approaches are difficult to understand and verify, which limits trust and wider use in safety-critical applications. This project will develop new methods that will help robots better understand human intent, provide timely assistance, and maintain safe and predictable performance, even in uncertain situations, allowing humans and robots to share control more effectively while reducing workload. The broader impacts of this project will include more trustworthy and reliable robotic systems for healthcare, manufacturing, and exploration in space and the ocean. The project will also promote education in science, technology, engineering, and mathematics by providing hands-on research opportunities and training for undergraduate and graduate students, engaging K–12 students in interactive robotics activities, and collaborating with national laboratories and industry partners to establish safety standards and encourage the responsible use of advanced robotic systems. This project seeks to develop a unified framework for safe, reliable, and interpretable teleoperation, advancing the collaboration between humans and robots in high-risk, unstructured environments. The research integrates AI foundation models with formal methods for safety specification, enabling robotic systems to interpret human intent, adapt to dynamic conditions, and maintain predictable behavior even under uncertainty. Unlike existing teleoperation approaches, which purely rely on opaque AI “black-box” models or oversimplified safety assumptions, this work focuses on producing certifiable and generalizable methods that move beyond basic collision avoidance to address real-world complex safety requirements. The proposed research aims to advance three core capabilities. First, it will develop learning-based assistive primitives that connect operator intentions and generate reusable, specification-compliant actions with minimal training. Second, it will establish logic-guided motion prediction and control frameworks that incorporate uncertainty quantification to ensure safe human-robot collaboration. Third, it will implement real-time monitoring and safety mechanisms capable of detecting anomalies, generalizing to novel conditions, and maintaining situational awareness throughout teleoperation. Collectively, these innovations are designed to enable adaptive, trustworthy, and high-performance human-robot interaction in real-world scenarios, laying the foundation for broader deployment of teleoperated 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.

NIH Research Projects · FY 2026 · 2026-05

Project Summary: Molecular modeling and theoretical characterization of protein dynamics and ligand binding Molecular recognition plays a crucial role in biology, chemistry and medicine. To answer key biological questions and enhance drug development, it is essential to understand protein-ligand binding at atomic detail. While multiple factors (i.e. pharmacokinetics) influence drug efficacy, protein-drug recognition is essential for achieving effective therapeutic action. While recent advances in molecular modeling show great success in computing ligand-protein binding affinity (i.e., binding free energy ΔG), challenges remain, especially in protein systems with shallow ligand binging pocket and conformation changes induced by a binding partner(s). Moreover, drug candidates showing good in vitro binding affinity to their target protein are not always effective in cells. In some cases, binding kinetics are the major determinant of drugs’ in vivo efficacy. Yet, the nature of binding kinetics (kon and koff), especial drug binding residence time (~1/koff), is not well understood, preventing drug design for desired kinetic properties. In cells, many proteins are not a standalone protein and have binding partners, which can affect a target protein’s conformations/dynamics. Considering the protein complex can lead better drug design and open new opportunities to design “molecular glues” to degrade the target protein or strengthen its functional inhibition. Molecular dynamics (MD) simulation techniques are powerful tools to explore the dynamics of the molecular complexes. Still, there is incomplete knowledge of how/why the proteins fluctuate between multiple conformations in the local-protein complex environment and how to efficiently glue proteins with a small molecule. Our long-term goal is to bridge all these knowledge gaps. Here we continue developing and applying modeling tools to compute and understand drug-protein binding affinity and residence time. This research will also involve the ongoing development of artificial intelligence (AI) modeling to analyze the complex interactions and molecular motions to learn the physics that governs protein conformational transitions and ligand-protein binding. Using the findings from our deep learning (DL) models, we will use DL to further sample protein conformations and assist drug design. The proposed work expands our classical view of molecular recognition - knowledge of the free and final bound states of a ligand and a protein - to examining the hidden states unseen in experiments for computing drug binding ΔG and residence time to better translate in vitro binding studies to drug’s in vivo efficacy. We also consider target protein’s local-protein complex environment, allowing detailed molecular glue and PROTAC design and/or improved drug design. Three main overarching themes are: (1) efficiently compute and understand non-covalent binding residence time and affinity, (2) utilize protein’s binding partners to design highly specific drugs and develop molecular glues and PROTACs, (3) integrate physicochemical and AI modeling to understand governing physics for sampling protein conformations and drug design.

NIH Research Projects · FY 2026 · 2026-04

ABSTRACT: Dependence on host-cell machinery for protein synthesis poses a significant challenge to coronavirus replication, particularly at the onset of infection when viral genomic RNA must compete with an abundance of host mRNAs for translation. To overcome this hurdle, viruses have evolved sophisticated strategies to commandeer the host translation machinery. While multiple mechanisms by which SARS-CoV-2 hijacks host translation have been elucidated, almost all involve non-structural viral proteins. This raises a fundamental question: how does SARS- CoV-2 establish a translational foothold during the early stages of infection, before non-structural proteins are synthesized? The viral nucleocapsid (N) protein is the primary viral factor present at this early stage of infection and has been shown to manipulate cell machinery to facilitate infection. We have found that N protein physically and functionally interacts with the human translation machinery, facilitating preferential viral translation. Moreover, our results suggest that the viral genome's 5ʹ untranslated region exploits high-affinity N protein binding to potentiate selective viral RNA recognition for translation. We hypothesize that N protein is a key mediator of viral translational hijacking in early SARS-CoV-2 infection, establishing a new paradigm within the N- protein functional repertoire. This proposal now seeks to elucidate the molecular mechanisms of host protein- synthesis modulation by N protein for viral benefit during early infection. Through two specific aims, we will (1) identify the viral determinants responsible for the impact of N protein on viral RNA translation and (2) delineate the roles of host factors in N protein viral-translation enhancement. By combining biochemical, biophysical, and genetic approaches, we will establish a comprehensive understanding of unanticipated host-virus interactions that govern SARS-CoV-2 pathogenesis, uncovering novel viral vulnerabilities that can be exploited to develop targeted antiviral therapy. Ultimately, this study will provide new insights for innovative therapeutic strategies that can be extended to other viruses with RNA-binding proteins, offering a promising avenue for smothering infection at its onset. 3

NIH Research Projects · FY 2026 · 2026-04

Project Summary Environmental exposure can alter pseudouridine (Ψ) landscape in cellular RNA. Ψ is the most abundant modified nucleoside in nature and it regulates many RNA processes. Similar to N6-methyladenosine (m6A), whose functions in RNA are modulated in part by its reader proteins, we reason that environmental exposure-mediated changes in RNA pseudouridylation and the resulting changes in RNA-protein interactions constitute a major regulatory mechanism of RNA processes that may underlie disease pathogenesis. Yet up to now, few Ψ reader proteins are known, and our knowledge is limited in the mechanisms through which environmental exposure- mediated aberrant RNA pseudouridylation and the ensuing perturbations of RNA-protein interactions contribute to the pathogenesis of neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS). Our proposed study aims to bridge this knowledge gap by uncovering the mechanistic link between environmental exposure and ALS pathology through the lens of RNA pseudouridylation and Ψ-protein interactions. Our preliminary data showed that: (1) at physiological concentration, profilin-1 (PFN1) binds directly and preferentially to Ψ-containing RNA; (2) ALS-linked point mutations in an ALS-associated protein, i.e., PFN1 reduce its affinity for Ψ-containing RNA, and (3) Ψ modification reduces RNA affinity for another ALS-associated protein. We organize our proposed research into three specific aims: Aim #1, to examine how RNA pseudouridylation is modulated by xenobiotic exposure in human neuronal cells; Aim #2, to investigate the impacts of RNA pseudouridylation on the transcriptome-wide occupancy of an ALS-associated protein; and Aim #3, to explore the impacts of RNA pseudouridylation and Ψ-protein interactions in xenobiotic and ALS pathological contexts. To accomplish these research aims and facilitate my transition to an independent academic career, I have assembled a mentoring committee with outstanding expertise in environmental toxicology, bioinformatics and neurobiology. The proposed research is built upon my extensive research skill sets, strong preliminary data, and the proposed mentoring/training plan. The outcome of the proposed research will unveil how environmental exposure-elicited aberrant RNA pseudouridylation alters RNA-protein interactions and contributes to ALS, thereby potentially leading to new biomarkers and therapeutic interventions. Moreover, the proposed career development and research plans will bridge gaps in my training and collect robust data for my independent publications and research grant applications, thereby transitioning me to an independent career at the intersection of environmental health science and neurodegeneration.

NIH Research Projects · FY 2026 · 2026-03

Summary This project seeks to develop enabling technology for the characterization of all human proteoforms. Proteoforms are the active species in biology and include not just the genetically encoded protein sequence but any alterations including alternate splicing, post-translational modifications, spontaneous chemical modifications, etc. As such their structure cannot be readily inferred from genomic data. Identification and characterization of proteoforms is the first step towards developing a complete mechanistic understanding of the cellular processes underlying both health and disease. Proteoforms are large, complex molecules that can be difficult to disambiguate. For example, an important functional difference could result from changing the location of a single modification, creating isomers that would appear to be nearly identical in an analytical sense. However, differences in key areas would distinguish these two proteoforms, and this proposal will develop methods to target those key differences by rationally and specifically fragmenting proteoforms at desired residues inside a mass spectrometer. This approach is an improvement over existing methods where information content is dependent on random fragmentation events, which may or may not reveal the key areas that differ. To enable rational protein dissection in the gas phase, chemical modifications must first be installed to create a weak link in the backbone chain that can be specifically triggered. The proposal will develop novel photoactivated modifiers that will selectively create radicals that will facilitate selective dissociation. Both the synthetic modifications and subsequent mass spectrometry are crucial innovations that will result from this research and further advance our ability to characterize proteoforms with the end goal of populating a human proteoform atlas.

NIH Research Projects · FY 2026 · 2026-02

Project Summary/Abstract Proteoforms are molecular forms of proteins in cells and tissues containing site-specific sequence variations and post-translational modifications. Proteoforms effectively describe cell phenotypes and provide important implications in disease mechanisms. Recent studies in immunobiology and disease pathology have emphasized intact proteoform characterization without enzymatic digestion in bottom-up approaches. Current mass-spectrometry-based top-down proteomics technologies fall short in capturing the complex proteome in small biological samples such as single cells. This proposal generates a suite of novel technologies for high-throughput omics- scale single cell proteoform profiling using innovative instrumentation, bioinformatics and high- throughput strategies. I will use human kidney cells as a model biological system in this proposal, and the approaches are generalizable to different cell types. I have recently developed a technology employing localized proteoform sampling coupled to single-molecule mass spectrometry to directly image and identify intact proteoforms in tissue sections (Su et al., 2022, Sci. Adv.). In Aim 1, I will expand the proteome detection and identification capabilities in this technology using innovative instrumentation, sampling and data acquisition algorithms. Together these will curate a knowledgebase serving as a proteoform library for kidney single cell proteoform profiling. In Aim 2, I will develop a novel platform leveraged for profiling of single cells dissociated from human kidney biopsies. I will develop a single cell preparation protocol for maximizing proteoform detection in kidney parenchymal and immune cells. I will also develop a bioinformatics approach tailored for proteoform identification in single cell datatype and discover proteoform signatures that differentiate cell types. In Aim 3, I will address the limitation in rare cell profiling by developing a series of high-throughput strategies including high-speed sampling and microarray cell patterning. These new technologies will unravel proteoform landscapes and signatures in rare kidney-resident innate immune cells (e.g., macrophages and dendritic cells) for the discovery of new cell populations that can be used as therapeutic targets and diagnostic tools for inflammatory diseases. My mentoring team consists of Dr. Neil Kelleher (mentor), a world-renowned protein biochemist, and Dr. Satish Nadig (co-mentor), a leading expert in kidney immunobiology. The proposed research is a substantial technological advancement in single-cell proteomics and sets a solid foundation for the pursuit of my independent career. The proposed research also well aligns with my long term research interest in developing enabling analytical technologies for biomedical science community with a special interest in the human innate immune system.

NSF Awards · FY 2026 · 2026-01

This project will create a new class of computational tools for physiologically accurate human motion simulation by bridging the critical gap between computer graphics and biomechanics. Simulation methods in computer graphics have historically prioritized computational efficiency and visually compelling results for animations and virtual experiences, often at the expense of physical accuracy. Conversely, biomechanical simulations emphasize realism and experimental validation but tend to be slower, more specialized, and less adaptable to interactive applications. By combining the strengths of both fields, the project will result in simulation methods that are fast, general-purpose, and physiologically grounded. This work will open the door to new cross-disciplinary collaborations, providing movement scientists in fields such as sports, health, and rehabilitation with tools to simulate complex, real-world movements that were previously infeasible. The resulting validated models can enhance training simulators and, when combined with existing open-source physics engines, will create new avenues for high-fidelity simulation and modeling in applications ranging from robotics to gaming. This project will deliver a next-generation, open-source physics simulator that accurately models musculotendon dynamics for graphics and other fields. To achieve this, the project explores four research thrusts. The first thrust establishes a unified, constraint-based simulation framework that treats muscles, tendons, skeletal structures, and environmental contacts as a coupled, fully implicit system. This formulation enables stable and accurate simulation of complex, high-contact human motion while maintaining physiological realism. The second thrust addresses muscle-based control by leveraging reinforcement learning to train neuromuscular controllers that produce realistic activation patterns, improving upon traditional joint-actuation systems/models that often generate unnatural and/or biomechanically implausible motions. The third thrust focuses on validation, using in-vivo biomechanical and physiological data, as well as benchmarking against existing simulation tools, to evaluate both the accuracy and computational performance of the system. The fourth thrust demonstrates broad applicability by enabling physiologically informed animation, injury-aware motion planning, and optimization of complex, contact-rich tasks. The resulting simulation platform is expected to support research and development across disciplines, contributing to improved understanding of human movement, better tools for clinical and biomechanical analysis, and enhanced realism in interactive 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 · 2025-11

Emissions from diesel engines negatively impact human health. Small diesel particulates (< 2.5 µm) have been linked to premature cardiovascular and respiratory deaths in metropolitan areas, as well as lung cancer. This project will investigate a new approach for electrostatic precipitation (ESP) technologies to reduce the emission of diesel particulates. The team will explore nanosecond high-voltage pulses to enhance ESP, also known as Plasma-enhanced Electrostatic Precipitation (PE-ESP). If successful, the new electrostatic precipitators will have a much smaller footprint. The new technology will open up new applications, such as in ships and trucks. By enabling cleaner transportation and shipping, the proposed work directly addresses urgent local and national air quality concerns, supports public health, and advances strategic efforts to meet stricter emissions standards and policy targets. This project will explore the application of nanosecond high-voltage pulse discharges as a novel approach in the context of electrostatic precipitation. The team’s preliminary results show that these nanosecond high-voltage pulses provide significant enhancement over conventional electrostatic precipitators (ESPs). However, the fundamental mechanism(s) underlying this enhancement are poorly understood. The fundamental understanding gained by this study will provide useful information about how to overcome current limitations and further improve PE-ESP. The studies include 1) investigating a reverse polarity two-terminal PE-ESP; 2) performing time-domain ESP simulations; and 3) evaluating a novel three-terminal PE-ESP configuration. The work is interdisciplinary, involving high voltage electronics, electrostatics, and fluid-dynamics, as well as combustion and aerosol science. The project will broaden its impact by expanding a workshop for high school science teachers, targeting schools in central Los Angeles and near the port area, to raise awareness about air pollution and engage students in scientific research. Undergraduate students will gain hands-on research experience to build confidence and interest in STEM careers. Additionally, a new course module on plasma-driven pollution remediation will integrate research findings directly into interdisciplinary curriculum development. 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

Abstract Title: Resolving competing light-to-electricity conversion mechanisms in nano-photodetectors for improved optoelectronic device performance Efficient photodetectors, essential components in technologies like telecommunications, medical imaging, and environmental monitoring, often face limitations due to competing electrical and thermal effects at very small scales. At the nanoscale, these two effects—the photovoltaic (PV) effect, where light directly generates electrical current, and the photothermoelectric (PTE) effect, where heat from light generates current—frequently interfere with each other, limiting the performance of devices such as cameras, sensors, and communication systems. This research aims to solve this fundamental problem by developing an advanced imaging method to precisely map and distinguish these effects at scales smaller than the wavelength of visible light. The outcomes of this research will significantly improve photodetector efficiency, enabling faster telecommunications, better medical diagnostics, and improved environmental sensing. Additionally, the project will support education and training for students through hands-on activities, workshops, and research experiences, preparing them for careers in rapidly evolving fields like nanotechnology and photonics. The technical objective of this research is to understand and control the interplay between PV and PTE effects in low-dimensional nanostructures to optimize photodetector performance. The research introduces a novel 3D Near-Field Scanning Photocurrent Microscopy (3D-NF-SPCM) method with sub-5 nm spatial resolution to disentangle these effects. The approach is organized into three synergistic research thrusts: (1) engineering thermal and Seebeck gradients in two-dimensional heterostructures (e.g., graphene/h-BN, carbon nanotube/MoS₂ hybrids) to systematically control PTE effects; (2) using nanoscale defects and strain gradients to decouple and coherently couple PV and PTE mechanisms, optimizing their combined response; and (3) integrating these nanoscale insights into real-world device architectures through plasmonic slot waveguides. The project will address current limitations in characterizing nanoscale optoelectronic properties, provide predictive design rules for high-performance photodetectors, and deepen fundamental understanding of nanoscale light-matter interactions. Results from this work will guide future developments in integrated optoelectronics, quantum sensing technologies, and nanoscale energy conversion 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 · 2025-10

Infants improve their motor skills, such as sitting, crawling, and standing, through practicing these skills in everyday life. However, limited tools for measuring infants’ movements in the home are a barrier to understanding the factors that give rise to individual differences in opportunities for movement as well as individual differences in caregiving practices that can predict motor learning outcomes in infancy. This project uses new wearable sensor technologies combined with Artificial Intelligence to record infants’ behavior across a week-long period to understand how the patterns of infants’ movements unfold over time. Results of this research advance understanding of motor development and provide useful information for clinicians to help promote healthy motor development in infancy. This project aims to collect data from families across the United States by mailing wearable sensors for infants to wear over the course of a week. In contrast to current methods that rely on labor-intensive manual coding of infant movement data, this project leverages Artificial Intelligence to (1) measure the amount and types of movement that 7-month old infants engage in at home during a typical week, (2) examine caregiving practices that give rise to individual differences in opportunities for movement, and (3) predict individual differences in motor development outcomes, including sitting and standing proficiency, at 11 months of age. Collecting and sharing a large longitudinal dataset of wearable sensor data enables advances in understanding motor development trajectories as well as advances in Artificial Intelligence methods for robustly identifying human movement. 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

Computer programs today are essential to everything from smartphones and vehicles to national defense. However, the software that runs these systems is often large and complex, making it difficult to detect hidden problems before they cause real-world harm. Existing tools used to find flaws, such as bugs or vulnerabilities, can either miss important issues, report many false alarms, or take too long to run. This project will investigate how artificial intelligence, specifically Large Language Models (LLMs), which are systems trained to understand and generate text (including code), can assist in the analysis of software. Instead of using LLMs by themselves, this project explores how they can support existing analysis methods and tools, improving their accuracy and speed. The successful completion of the proposed activity will lead to changes in the way programs are analyzed to find bugs and vulnerabilities. The project hypothesizes that LLMs can be used as a complementary approach to conventional program analysis by combining the strengths of both to address key challenges in analyzing large and complex software. The research will begin by identifying the limitations of current program analysis techniques, such as the tradeoffs they must make between precision and scalability. It will then explore strategies for using LLMs in supportive roles to target these limitations and assist with particularly difficult aspects of program analysis. The design space for integrating LLMs with existing tools is broad and includes possibilities such as bug-type-specific workflows, autonomous agents, and mechanisms for verifying the correctness of LLM-generated results. The project will evaluate these designs in real-world applications, including tools for vulnerability discovery and enhanced operating system testing. The expected outcome is a set of more effective and practical analysis tools that advance both software security and the understanding of how to combine artificial intelligence with program analysis methods. 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

Despite the growing need for spectrum resources, spectrum licensed for commercial use remains largely underutilized due to lack of established market mechanisms for dynamic trading of the spectrum between cellular operators. Further, there are no protocols that implement such sharing and exchange on short time scales. Multi-operator spectrum sharing is therefore necessary not only to make best use of this critical resource (spectrum), but also to improve the profitability of spectrum service providers and reduce the cost to customers. Towards this goal, the project aims to develop new market mechanisms and protocol support for dynamic and automated spectrum sharing between cellular providers. Deriving motivation from electricity markets, the project analyzes a two-step design for spectrum markets, involving trading of both forward and spot spectrum contracts between cellular operators. Building upon the Open Radio Access Network (O-RAN) software framework, the project also analyzes designs that provide logical connectivity from different radio access networks (that might belong to different operators) to the cellular core networks of one or more of the chosen operators. This project explores a design of a two-timescale market involving a forward spectrum market (FSM) and a spot spectrum market (SSM) through which forward and spot spectrum contracts are traded between cellular providers, in addition to any bilateral settlements that may exist over longer timescales. This enables flexible sharing of radio resources between cellular operators. Further, this project utilizes network slicing within the framework of the O-RAN architecture and protocols to realize the forward and spot contracts in a secure and efficient manner. Different from traditional roaming, network slices seek to provide users of the slice an assured amount of bandwidth and latency through service-level agreements (SLAs). The project also seeks to implement prototypes of the market solution and the network slice implementation over cellular networks, where the market clearing solutions are fed to the 5G core network that supports network slicing. The broader impacts of the project are realized through continued collaboration with cellular network operators and device vendors, incorporating research insights into courses and capstone projects, and involving undergraduate student researchers in developing a Spectrum Trading Game aimed at motivating high-school students towards science and technology careers. 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

In the future, autonomous systems will increasingly rely on large networks of intelligent agents, such as drones, vehicles, robots, and smart infrastructures. These agents will need to work together and coordinate their actions. Achieving trustworthy and resilient coordination among collaborating agents is challenging, particularly when they may malfunction, act in opposition, or vary in capabilities and objectives. This project aims to revolutionize multi-agent systems (MAS) that exhibit resilient, adaptive, and trustworthy behavior by creating intelligent coordination mechanisms capable of maintaining performance and safety even in the presence of uncertainty, failure, or conflict. The expected outcomes of this research will lay the groundwork for building autonomous systems that can be confidently deployed in complex, real-world scenarios. The research will advance foundational knowledge through three key efforts: (1) designing algorithms to detect and mitigate abnormal behaviors in cooperating agents; (2) managing adversarial or non-cooperative agents using game-theoretic and adversarial machine learning methods; and (3) enabling resilient coordination among diverse agents through robust distributed control frameworks. The proposed work will support high-impact applications in transportation, disaster response, and smart infrastructure, where reliable MAS coordination is critical for public safety and operational efficiency. It will inform best practices and ethical guidelines for integrating AI and multi-agent systems into critical infrastructure, ensuring fairness, transparency, and reliability in their deployment, and ultimately foster public trust in AI technologies and contribute to sustainability, safety, and social good. This research will also strengthen research capacity by expanding interdisciplinary collaborations, developing new curricula and workshops, and offering hands-on research opportunities for students from a wide range of backgrounds. 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

Modern imaging technologies are central to progress in science, medicine, and engineering. Yet, many advanced imaging systems operate under physical or resource limitations that make it difficult to directly acquire high-quality images. Computational imaging addresses these challenges by using algorithms to reconstruct images from incomplete or indirect measurements. In recent years, deep learning has enabled new capabilities in computational imaging, but current methods assume that the training and test data share the same conditions. This assumption often does not hold in real-world settings. This project addresses this critical gap by developing new methods to ensure that deep-learning models for image reconstruction remain reliable and accurate even when the data conditions shift. The outcomes of this research will have broad use and transformative effects across a wide range of scientific, engineering, and biomedical applications, where robust image reconstruction is essential. Broader-impact activities include the organization of special sessions, workshops, and journal issues for the computational-imaging community, dissemination via open-source code, and curriculum development at both institutions. This project focuses on score-based models — a class of deep generative models that solve imaging problems by learning the score function of the image distribution. The central goal is to develop a unified mathematical framework for analyzing and improving the robustness of these models under distribution shifts between training and test data. The project introduces Robust Score-based Inversion (RoSI) as a foundation for (i) quantifying the extent of such shifts using the model's own score function; (ii) characterizing the effect of shifts on reconstruction and sampling performance; and (iii) enabling principled adaptation of models to new imaging settings. The research will be validated in real-world imaging systems, including lensless cameras, computational microscopes, and magnetic resonance imaging, providing both theoretical insights and practical tools for reliable computational imaging. The project will also promote education and engagement in the areas of computational imaging and machine learning. 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

With support from the Improving Undergraduate STEM Education: Hispanic-Serving Institutions (HSI) Program, this ELPSE Level 2 project aims to expand access to artificial intelligence (AI) education through the development of interdisciplinary AI minor and certificate programs, along with community-engaged research opportunities. The project will serve students in the Inland Empire region of Southern California, where there is strong demand for skilled workers in computing and AI-related fields. Building on ongoing collaboration between California State University, San Bernardino, and the University of California, Riverside, the initiative will provide students with flexible entry points into AI careers through curriculum development, industry-informed research experiences, and faculty mentoring. The program will be open to all students across disciplines and institutions. The specific aims of the project are to: (1) build new AI minor and certificate programs that provide students with theoretical and applied skills through career-relevant curriculum design and project-based coursework; (2) establish an AI Help Desk to support community organizations and small businesses by offering AI consultations and student-led solutions to real-world problems; (3) engage students from a range of disciplines, including non-computing majors, through mentoring, faculty-guided research, and community workshops; and (4) sustain pathways into AI by integrating formal instruction with informal learning and collaborative research opportunities. The project will investigate how applied AI learning experiences influence student outcomes and motivation. A mixed-methods research design will be used, combining survey data, academic performance metrics, interviews, and focus groups to assess outcomes. Faculty teams will also evaluate the impact of curriculum design and community-based projects on student engagement and learning. Results will be shared through academic publications, open-source tools, presentations at education and computing conferences, and local institutional partnerships. The long-term goal is to develop a scalable, replicable model for AI education that aligns with regional workforce needs by offering flexible academic programs and practical, hands-on learning experiences to strengthen AI readiness and support sustained growth in the STEM workforce. This project is funded by the HSI Program, which aims to enhance undergraduate STEM education and increase capacity to engage in the development and implementation of innovations to improve STEM teaching and learning at HSIs. 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 award is funded in whole or in part under the American Rescue Plan Act of 2021 (Public Law 117-2). Most cells in our body have a primary cilium (a single hair-like projection on their cell surface) that allows them to interact with each other via chemical signals. This ancient and mysterious organelle acts like a radio antenna, allowing cells to receive different information, depending on which kinds of receptors their antenna contains. Gene mutations affecting ciliary signaling have been implicated in a host of diseases that result in the malformation and dysfunction of multiple organs. This project focuses on primary cilia in the developing brain. During brain formation, the progenitor cells that give rise to nerve cells (neurons) coordinate with each other to make sure that specific type of neurons are produced in the correct location at the right time. Surprisingly little is known about which receptors in their primary cilia allow effective communication. This project will document the full catalog of receptors situated in the primary cilia of neural progenitor cells using novel techniques that selectively label cilium proteins with a special tag named biotin, and then identify these marked proteins using mass spectrometry. It will also decipher the mechanisms by which some of these proteins influence developmental decisions. Pilot studies have already identified proteins previously shown to control brain development that are unexpectedly operating in the cilium. Overall, the results of this project will provide novel insights into how neural cell progenitors coordinate to build the brain. They advance our understanding of how brain development works and will also help us to better understand brain developmental disorders. In addition, this project will engage local K-12 students and high school teachers in laboratory research, aiming to inspire a sustained interest in the biological sciences among local youth, as well as implementing an integrated educational program to promote participation and retention of students in biological research. These activities will substantially improve education in life science in the California Central Valley. Neural progenitors in the developing brain (also known as Radial Glia, RG) produce all of the brain’s neurons in a time- and space-specific manner. Different neuronal types are generated at distinct developmental stages and in discrete brain regions. This process is highly coordinated via cell signaling sensed by the RG primary cilia. Defects in cilium function lead to ciliopathies, a wide-ranging spectrum of disorders that usually involve brain structural defects. Yet a systematic understanding of RG cilium signaling pathways is lacking. This project will identify new signaling proteins in RG cilia by leveraging a new proximity labeling tool (TurboID) and a novel transgenic mouse model in which TurboID is selectively expressed in the cilium of RG cells. Quantitative proteomic studies with rigorous controls will reveal bona fide cilium proteins operating in discrete brain regions and across different developmental stages. Pilot studies with cilium-targeted TurboID have surprisingly revealed cilium localization of a protein previously reported to regulate neurogenesis through undetermined mechanisms. Preliminary data show that this protein operates in the cilium to regulate Hedgehog signaling, the best-described pathway in the cilium. Investigating the interaction between this new cilium protein and Hedgehog signaling will reveal new regulatory mechanisms in embryonic neurogenesis, and demonstrate how cilium proteomics can help solve standing long-questions in brain development. By systemically unveiling the signaling pathways used by RG cilia with spatiotemporal resolution, this project will also chart new directions for future neurodevelopmental studies. 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

Despite the advances in deep neural networks (DNNs) and edge computing, there exist substantial challenges to enabling end-to-end DNN inference on a full spectrum of edge devices, such as tiny wearables and low-cost Internet-of-things (IoT) devices. This problem has spurred the recent studies of brain-inspired vector symbolic representation (VSA) classifiers as an alternative framework for ubiquitous on-device inference. At a high level, VSA classifiers mimic the brain cognition process by representing each object as a vector (typically in a very high-dimensional space). While VSA classifiers offer advantages over DNNs in terms of inference efficiency due to parallel processing, the hyper-dimensionality in their design can still easily result in a prohibitively large VSA model size beyond the limit of many tiny devices with stringent resource constraints. If successful, this project will make it possible for more everyday devices to run advanced artificial intelligence (AI) on their own, without needing to send data to remote servers. This could improve privacy, save energy, and open the door to smarter wearables, medical devices, and home gadgets. Finally, the project will bring the latest discoveries into college courses to help train the next generation of engineers and computer scientists. To address the hyper-dimensionality challenge, this project moves away from hypervector-oriented VSA and proposes TinyVSA, which uses much smaller, compact vector representations. Specifically, this project focuses on three key directions: first, redesigning TinyVSA’s vectors to improve accuracy while the VSA dimensionality by orders of magnitude; second, making TinyVSA run continuously and efficiently on tiny, low-power chips; and third, developing an efficient, hardware-aware method to automatically find the best TinyVSA architecture for target devices. 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

A diode laser device is made from semiconductor, a material with its electrical conductivity dictated by the quantities of negatively and positively charged mobile particles within itself, which are referred as electrons and holes, respectively. An n-type semiconductor possesses more electrons than holes while a p-type semiconductor has more holes than electrons. Traditional diode laser devices contain a key component equivalent to a p-n junction, where electrons and holes meet and recombine to emit light. Diode lasers with emission wavelengths in the infrared and visible spectral bands are widely commercially available. In contrast, those devices emitting lasing with wavelengths less than 315 nanometers in the deep ultraviolet (UV) spectral bands (UV-B and UV-C) are severely underdeveloped. This is because the material used to develop those lasers, namely aluminum gallium nitride (AlGaN), has fundamental issues including weak p- and n-type conductivities. This project addresses this challenge by developing novel metal-semiconductor-metal random laser devices based on magnesium zinc oxide (MgZnO) semiconductors. These devices use the injection of high-energy electrons from power supply to generate amplified numbers of electron-hole pairs for lasing, thus circumventing the strong p-type requirement in conventional semiconductor p-n junction lasers. The principal investigator and his students will work to achieve MgZnO lasers with deep-UV emission wavelengths between 315 and 200 nanometers. In addition, a novel approach will be used to convert random lasers which emit light with multiple wavelengths in all directions into highly directional, single-wavelength MgZnO lasers. The success of this effort will enable the development of portable semiconductor lasers in the deep-UV spectral bands for next-generation data storage and recording, medical diagnosis and surgery, photodynamic therapy, biological agent detection and sterilization, and water purification. Thus, it will have a profound positive impact on national health, prosperity and welfare. This project will train PhD students and undergraduate students who will graduate with versatile skills to advance semiconductor photonics technology in industry, academia, or government. Educational outreach will be extended to K-12 students to foster their interest in semiconductors. Additionally, the project will help train semiconductor nanotechnology technicians, a workforce increasingly in demand by society. Technical Description: This project seeks to overcome fundamental challenges to the development of deep-UV semiconductor lasers. It will demonstrate novel MgZnO deep-UV laser devices with an emission wavelength range between 200 and 315 nm in the UV-B and UV-C bands using metal-semiconductor-metal junctions rather than conventional p-n junctions. Gallium-doped n-type MgZnO nano-column semiconductor thin films with various Mg mole fractions will be grown using molecular beam epitaxy. Deep-UV metal-semiconductor-metal random lasers with scalable output power as a result of impact ionization, and their mode-coupled highly directional and single-mode lasers will be simulated, fabricated and characterized. Despite the enormous success of semiconductor lasers in all other spectral ranges including near-UV, visible, infrared and terahertz, electrically driven coherent single-mode semiconductor lasers operating at wavelengths less than 315 nm at room temperature are rare. This research will fill the 200 to 315 nm range wavelength gap, which will impact areas such as information storage, display and imaging, biology and medical therapeutics, and water purification. 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

Since graphs can readily represent entities and relationships among them, they are widely used to represent large volumes of data from domains ranging from transportation networks to biological networks. When the data has an associated temporal dimension, it is represented as an evolving graph that consists of a sequence of snapshots of the graph at different points in time. Mining of large evolving graphs involves understanding trends in changes to relevant graph properties over a chosen time window. Since evolving graphs are extremely large, evaluating queries over a sequence of snapshots is both compute- and data-intensive. The irregular structure of real-world graphs and the iterative nature of graph queries that require multiple passes over graph data impose further challenges to optimizing the evaluation of temporal queries. This project aims to dramatically improve parallel evaluation times and memory requirements of evaluating temporal queries on evolving graphs. Building a powerful system will accelerate discoveries in fields that employ evolving graph analytics. In addition, it will result in training graduate students in high-performance computing, an area of national need. The software and graph data developed during this project will be available to other researchers. The technical aims of this project are to substantially advance the state of the art of evolving graph analytics by developing highly scalable systems and to expand the scope of supported analytics queries greatly. For graphs that have been evolving over a long period of time, two classes of challenges are addressed. The first class requires substantially improving the efficiency of evolving graph processing. The large sizes of graphs and a large number of snapshots lead to high query evaluation costs. Novel approaches that exploit the slowly changing nature of an evolving graph are needed to address these challenges. The second class aims to carry out relevant snapshot identification and planning problems. Relevant snapshots identification omits snapshots that do not contribute to property change, while planning considers goal-oriented evolution of the graph to meet future needs. 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.

NIH Research Projects · FY 2025 · 2025-09

Project Summary/Abstract Cells are the basic unit of life. Cells are very dynamic: they change over time and locations, respond to different environments, and interact with other cells. Over the past decade, single-cell biology has witnessed enormous growth owing to massive technical advances, such as single-cell sequencing, multi-omics, and spatial omics. However, obtaining dynamic dimensions of live cells along with their multi-omic information at the single-cell resolution is currently difficult and certainly not possible on large scales. Here, we propose a novel cell barcoding technology that has the potential to enable us to collect live information of cells and connect the data to their detailed omics information. This technology makes use of laser particles (LPs) with unique optical barcodes for >100,000 channels, each containing a unique DNA barcode. The “dual-barcoding” will allow us to optically track live cells under a microscope while they are in their natural environment or in culture, acquire their live information, harvest the cells, acquire the omics information of the same cells by droplet-based next-generation single-cell sequencing, and then combine the live imaging and omics data at the single cell resolution. Furthermore, our technique can be upgraded to multi-omics modalities, combining multiple layers of information from the genome, epigenome, transcriptome, and proteome, together with morphological, locational, functional, and behavioral data. We will apply the method to study sentinel lymph node (SLN) metastasis of cancer cells in vivo. The acquired in vivo single-cell imaging and multi-omics data will provide an unprecedented picture of the cancer cell lymphatic metastasis process. This project has two specific aims. Aim 1 will develop an optical-and- DNA “dual” barcoding strategy for droplet-based single-cell sequencing. Aim 2 will apply the method to study breast cancer SLN metastasis in vivo. During the K99 period, the applicant will receive additional training to expand her experience and shape her independence in the following areas: (1) LPs and optical barcoding, (2) LP imaging and in vivo mouse imaging, and (3) single-cell sequencing and multi-omics. This proposal is under the combined mentorship of Dr. Andy Yun (LP technology, optics, and imaging) and Dr. Ralph Weissleder (cancer biology, in vivo imaging, and system biology), and a team of experts as advisors for single-cell sequencing and bioinformatics. The interdisciplinary research environment at Massachusetts General Hospital and Harvard Medical School will significantly facilitate the proposed study. If successful, the proposed study will offer a new paradigm for “dynamic” single-cell analysis, with unprecedented speed and throughput, enabling multi-omics modalities for the profiling of proteins, RNAs, and DNAs at the single-cell level, together with cells’ dynamic phenotype information, enable spatial-omics profiling at the 3D resolution without the need for cell segmentation. This will be a significant step beyond the current single-cell omics strategies that collect only snapshot data, in vitro or ex-vivo. This new method will transform the way we use imaging and single-cell analysis and will open enormous applications for scientific discovery, diagnosis, and treatment in healthcare.