New York University

NSF Awards · FY 2026 · 2026-09

Artificial intelligence is increasingly used to model complex scientific and engineering systems. Maintaining leadership in this area over the long term depends not only on larger models and more data but also on mathematical methods to enhance training and to make training more stable, reproducible, and mathematically well-founded. This project will develop mathematical principles for designing training objectives and methods for generative artificial intelligence that remain stable under realistic training conditions, such as finite data, rather than producing misleading updates or fragile models. This project directly advances artificial intelligence as an area of federal strategic interest by strengthening the foundations needed for reliable artificial intelligence models in scientific and engineering applications. More broadly, this project will strengthen the scientific and industrial artificial intelligence ecosystem in the United States by developing reliable methods for discovery, design, and decision-making in complex systems. The project will also support education and workforce development through graduate student training, integration of project ideas into courses, open source software, and public benchmark problems. The project will develop stable formulations and numerical discretizations of loss functions for generative artificial intelligence models of time-dependent stochastic processes. These models often use training objectives involving time and space derivatives, but current practice commonly estimates such objectives from samples in a black box manner, which can introduce systematic errors, poor conditioning, and unstable training. The project will show that loss functions for certain generative models can be interpreted as variational formulations of partial differential equations. This connection will make it possible to transfer concepts such as stability, coercivity, structure preservation, and consistent discretization into the design of empirical loss functions for data-driven generative modeling. The work will establish rigorous correspondences between continuous loss functions and partial differential equation formulations, derive well-conditioned discrete loss functions that remain stable when only data samples are available, and develop discretization strategies that also provide algorithmic benefits such as parallel training across time. The resulting methods will be applied to reduced and surrogate modeling of stochastic and chaotic systems, to demonstrate more stable, reliable, and efficient generative reduced models. 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

While formal physics education typically begins in high school or college, humans develop sophisticated intuitions about the physical world long before entering a classroom. Even young children can predict whether a tower of blocks will fall over, how much weight a tree branch can support, or how far a ball will roll if kicked. This intuition about the physical world is a core part of human intelligence, contributing to our everyday commonsense knowledge, but it is still quite difficult to engineer systems that match the robustness and performance of this kind of human intelligence and that would enable robots and other machines to safely and intelligently interact with the real physical world. The current project aims to “reverse engineer” this aspect of human intelligence to determine what internal processes people are using to reason about the physical world. This project relies on a custom physics simulator that can express a wide range of possible physics, many of which differ from those experienced on Earth (e.g., gravity that is a little too strong or an unnatural relationship between force and motion). Human participants and AI machines will judge which physical laws seem most natural or correct and infer unseen parts of a scene using only the motion of visible elements. The comparison of human and machine strategies on identical tasks will uncover the representational commitments of each while advancing state-of-the-art methods for evaluating AI systems and providing inspiration for design of next-generation AI systems. The interdisciplinary nature of the project fosters workforce development at the cognitive science-AI frontier. Open-source tools, including physics engines, data analysis protocols, and datasets, will be freely accessible to the broader research community. This project advances knowledge about human intelligence and addresses NSF strategic priorities in AI research and understanding human intelligence and the White House’s “AI Action Plan” emphasis on preserving American leadership in transformative AI and maintaining global competitiveness in scientific domains of artificial intelligence, cognitive science, and education. 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

Understanding how our bodies fight disease is essential to advancing medicine and increasing the quality of life. Modern technologies, known as spatial profiling, allow us to scan human tissue to derive genetic and physical information about biological processes. But the resulting data is so large and complex that even experts cannot fully explore it with the naked eye. This project will yield novel human-guided artificial intelligence (AI) approaches to steer and interpret artificial intelligence through interactive visual interfaces. The technology will assist researchers in analyzing biological phenomena, including the causes of diseases (notably cancer), immune responses, and the effects of therapies at the cellular level. Since these phenomena rely on the spatial and temporal organization of cells within the tissue, we will develop an interactive biomedical atlas with embedded analytics to visually analyze these structures and communicate scientific findings to other researchers, practitioners, and the general public. A joint educational plan pairs computer science and biology students to learn and build the research tools of the future. Through these efforts, we aim to revolutionize our understanding of human disease and increase life expectancy for all Americans. Lessons learned in human-AI collaboration and spatial data analysis can expand to other disciplines, including earth and space science. This project develops novel human-in-the-loop visual analytics approaches for spatial-profiling data. By combining imaging with spatially referenced data, the project can map complex multimodal data with greater precision. However, the immense scale and complexity of the multi-modal data, paired with the explorative nature of pre-clinical research, make fully automated analysis unreliable. To bridge this gap, this project proposes interactive visualizations that let experts steer and interpret AI-driven insights, organized into four themes: (1) designing scalable algorithms for joint image and omics representation, (2) integrating machine learning into interactive workflows for expert-led spatial and temporal analysis, (3) creating narrative tools to disseminate scientific findings, and (4) unifying these tools into a single framework validated by real-world biomedical data. The outcomes of this research will drive precision medicine, increasing survival rates and improving public health. Advances in human-centered computing will enhance human-AI collaboration for analyzing imaging and omics data, with applications reaching beyond biomedicine into material, geographic, and astronomical sciences. Our educational initiative bridges human-centered computing and biology, launching new bioinformatics curricula to develop a multidisciplinary workforce at the undergraduate and graduate levels. 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

Understanding how genes are turned "on" or "off" (regulatory control) is central to biology, medicine, and biotechnology. Although the human genome has been extensively mapped, scientists still do not know whether the linear arrangement of DNA elements encodes an underlying design logic governing how groups of genes are activated together. This project tests the idea that the functional behavior and three-dimensional organization of the genome are encoded in its linear structure. By systematically re-engineering the spacing and organization of regulatory elements to control coordinated gene activity in human cells, the research aims to move beyond observing the genome toward the ability to predict and design its behavior. These advances have broad implications for the US national interest, including improving the reliability and potency of engineered cells for therapeutic applications, enabling more precise control of gene expression in regenerative medicine and immune engineering, and supporting the development of robust multi-gene systems for biotechnology and biomanufacturing. In parallel, the project integrates research with education through a tiered training program spanning middle school lab visits, undergraduate and master’s research, and graduate training, contributing to the development of a skilled and committed workforce in biotechnology and related fields. This project, SynTACS (Synthetic Transcriptional Architecture of Condensates and Super-enhancers), will test the hypothesis that the linear organization of super-enhancers encodes transcriptional control by shaping transcription factor condensate dynamics and chromatin topology. Using REWRITE, a platform for programmable locus-scale (>100 kb) DNA restructuring in human induced pluripotent stem cells, the project will systematically reconfigure the density, spacing, and orientation of regulatory elements within a 209 kilobase multi-gene region at its native genomic context. Engineered variants will be evaluated through three integrated approaches: (1) transcriptional profiling to quantify gene expression changes across constructs, (2) live-cell imaging of transcription factor condensates to measure their formation, dynamics, and stability, and (3) high-resolution chromatin topology mapping using region-capture Micro-C alongside epigenetic profiling. By linking genome architecture to molecular dynamics and functional output in a controlled, causal framework, this work will establish design principles governing coordinated gene regulation. The outcomes will provide a foundation for engineering synthetic multi-gene regulatory systems and advancing genome-scale biotechnology. 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 · 2026-06

Malaria presents a significant health challenge in sub-Saharan Africa (SSA). It is a threat to US citizens in relation to health security, safe travel, missionary work and military operations. The relevance of malaria control and elimination in SSA to US economic opportunity and role in geopolitics cannot be underestimated. In SSA, a large fraction of the population across all ages harbors the malaria parasite Plasmodium falciparum without clinical manifestation, providing a vast infection reservoir for mosquito transmission. This asymptomatic reservoir is sustained by enormous antigenic diversity of the parasite. Our proposal addresses the hypothesis that progress toward falciparum malaria elimination should focus on reducing the number of antigenically diverse strains. Here we test this hypothesis in Ghana, a stable country in West Africa with a National Strategic Plan for malaria elimination including regions of economic and strategic importance to the US. Malaria Reservoir Study-II (MRS-II) takes advantage of enhanced intervention efforts in northern Ghana as major perturbations reducing the transmission reservoir, to address the sharp transition postulated by previous theory in both parasite diversity and infection prevalence with changes in transmission intensity. We test the importance of strain diversity to resilience of the transmission system, with a framework that combines longitudinal surveys of unprecedented depth with novel metrics for molecular surveillance of variant antigen diversity and computational models encompassing generalized and variant antigen immunity. Our three major inter-related aims are: (1) Field work for surveillance of enhanced interventions in two sites of northern Ghana differing in recent malaria control; (2) Characterization of the resilience of high-transmission systems to the major perturbation of targeted mass drug administration to children aged 3 months to 12 years in terms of changes in strain diversity; (3) Computational modeling to evaluate and predict dynamical responses to enhanced intervention in the context of the fundamental positive feedback between strain diversity and transmission intensity. The project will provide real world experience for US trainees in disease surveillance and control as well as training in quantitative tools for biomedicine. Quantitative characterization based on molecular surveillance, together with mathematical models that consider strain diversity and structure, will inform public health policy on the impact of control efforts in high-transmission regions. Current efforts are designed to reduce disease burden in children with no surveillance of the impact on the reservoir of asymptomatic infections in the community. MRS-II fills this significant gap in regions of northern Ghana with relevance to high-transmission areas of the Sahel in West Africa. More generally, our study aims to shift the paradigm in malaria epidemiology to make the parasite rather than the infected host the unit of measurement for evaluation of control and a target for elimination.

NIH Research Projects · FY 2026 · 2026-06

Project Summary Fragile X Syndrome (FXS) is the leading monogenic cause of intellectual disability and is frequently associated with autism spectrum disorder. FXS is caused by the loss of expression of the Fragile X messenger ribonucleoprotein (FMRP) which is a well characterized RNA-binding protein known to be critical for the regulation of protein synthesis and synaptic plasticity. Traditional studies of FXS have focused on the neuronal mechanisms and consequences of altered protein synthesis, yet recent studies have suggested the importance of astrocytes in regulating FXS neuronal physiology. Studies in both human and mouse models of FXS demonstrate that culturing FXS neurons with control astrocytes can ameliorate FXS neuronal alterations while culture of control neurons with FXS astrocytes can induce FXS-like physiology. Interestingly, each of these studies demonstrate that the observed alterations can be induced by the respective astrocyte conditioned medium, suggesting a secretory molecule may be responsible for these alterations. These results suggest that astrocyte genotype controls neuronal phenotype in FXS, yet the mechanism and molecules underlying these alterations have not been comprehensively identified. This proposal will address two questions: 1) What are the underlying changes to astrocyte protein synthesis and secretion follow FMRP loss? 2) What are the co-culture induced astrocyte-dependent changes to neuronal protein synthesis and which astrocyte-derived molecules are responsible? To identify the consequences of FMRP loss on astrocyte protein synthesis and secretion this proposal aims to apply ribosome profiling and astrocyte-specific secreted proteomics. To identify the co-culture induced astrocyte-dependent neuronal changes in protein synthesis this proposal applies ribosome profiling of human neurons in four co-culture conditions to test the impact of cell autonomous versus non-cell autonomous FMRP loss. Finally, this proposal aims to identify the astrocyte-derived neuron-internalized molecules responsible for these alterations using astrocyte-specific secretome labeling in combination with neuronal proximity labeling. All together this work aims to undercover novel modes of co-regulation between neurons and astrocytes critical to our understanding of neurological disease. It is the hope that these findings provide novel avenues for innovative therapeutic approaches for the treatment of FXS and other neurodevelopmental disorders.

NSF Awards · FY 2026 · 2026-05

Recent advances in generative artificial intelligence (GenAI) have transformed the way digital images and videos are created, enabling the automated generation of highly realistic content. However, these powerful technologies require massive amounts of electricity and computer memory, leading to high operational costs and significant infrastructure requirements. While these systems consume vast resources, the human eye and brain can only process a limited amount of visual information at any given time. This project addresses the gap between the high computational cost and the actual visual quality experienced by a human viewer. By aligning the efficiency of computer systems with the limits of human vision, this work aims to create cost-effective and faster digital content generation tools. This research supports the national interest by promoting economic prosperity through reduced industrial expenses and by accelerating the development of larger-scale, more efficient generative artificial intelligence infrastructure. This project establishes a research framework to measure, model, and optimize the relationship between computational expense and human visual perception for emerging GenAI models. The research methodology involves the development of psychophysical studies and large-scale datasets to quantify how specific hardware costs correlate with perceived visual quality. The investigator will then create new probabilistic models and guided artificial intelligence frameworks designed to maximize output quality under strict power and memory constraints. The technical scope includes the creation of an integrated, shared computing module for adaptive resource allocation. A case study on urban planning, conducted in partnership with small businesses, will demonstrate the practical application of these resource-aware frameworks. By bridging the gap between computing efficiency and human perception, the award provides foundational data and tools for a new class of human-centered visual generation technologies. 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/ABSTRACT Despite advances in collecting large-scale behavioral datasets, our ability to gain insights into an individual’s learning and decision-making processes remains limited. This is particularly true for characterizing individual dif- ferences in task performance, or how behavior in psychological tasks relates to psychiatric symptoms. Progress towards this ambitious goal depends on computational models that formalize the relationships between behavioral observations, the underlying latent cognitive processes, and individual differences in behavior. Unfortunately, ex- isting modeling approaches are either too simple to handle the highly variable nature of behavior, or too complex to yield interpretable insights into the cognitive processes of interest. An approach combining flexibility and inter- pretability could transform our understanding of healthy decision-making and psychiatric conditions. This proposal addresses this critical need by developing a novel computational framework to model an individual’s learning and decision-making processes in a flexible and interpretable manner. The proposal focuses on reward learning due to its critical role in healthy and dysfunctional decision-making, as well as its prevalence in psychology. Critically, our approach captures behavioral idiosyncrasies in individual subjects, instead of focusing on group averages. To achieve this specificity without undue sacrifices in interpretability, our framework relies on two techniques: very small recurrent neural networks (RNNs) trained to imitate an individual’s behavior, and dynamical systems theory to interpret how the RNN converts observations into decisions. Our prior research shows these tiny RNNs predict individual choices more accurately than classical models while revealing complex, previously unobserved learning strategies. Preliminary analyses suggest this approach discovers relationships in strategy use across tasks and identifies distinct patterns of decision-making based on clinical diagnosis. The proposed work has two primary aims. First, we will validate the stability of individual differences across multiple decision-making tasks by relating subject-specific strategies across tasks. Second, we will relate cognitive processes to psychiatric symp- toms by examining how strategies vary with symptom severity. We will also predict psychiatric symptoms based on individual differences in strategies derived from the fitted RNN models. Both analyses will use a large dataset (N = 815) currently under acquisition in the research lab of co-investigator Dr. Catherine A. Hartley, which in- cludes data from three decision-making tasks and an array of psychiatric symptom assessments. Our approach is a novel integration of data-driven and theory-driven approaches for computational psychiatry, offering a frame- work that can benefit from large datasets while still providing theoretical insights. This ability to generate cognitive theories from data alone could accelerate the study of individual cognitive differences, and particularly benefit the study of mental health. Ultimately, this could lead to more precise diagnostic tools and targeted interventions for psychiatric conditions by providing deeper insights into the cognitive mechanisms underlying decision-making.

NIH Research Projects · FY 2026 · 2026-04

Project Summary Neuropsychiatric disorders frequently fail to respond to first-line treatments, highlighting the urgent need for innovative therapeutic approaches targeting underlying neural mechanisms. A common feature across neuropsychiatric disorders is pathological expectation states, including both pathologically negative expectations (e.g., major depression: pessimistic future outlook, social anxiety: anticipated rejection, generalized anxiety: catastrophic forecasting) and positive expectations (e.g., gambling disorder: anticipation of unlikely wins, substance use disorder: overvaluation of drug rewards, bipolar mania: unrealistic optimism about risky behaviors). The proposed project will examine the causal role of frontal cortical circuits in generating and updating affective expectations. Leveraging cutting-edge neuromodulatory and high-density recording techniques, this research will determine how non-human primate orbitofrontal cortex (OFC) influences behavior through outcome expectation signals and how these signals can be modulated with precisely targeted electrical stimulation. The experimental approach centers on a novel behavioral paradigm that captures the dynamic nature of expectation-guided behavior in macaque monkeys. By monitoring continuous behavior following expectation violations, this paradigm provides unprecedented sensitivity to detect both the persistence of expectation bias and the temporal dynamics of behavioral updating. This investigation will: (1) Establish the causal relationship between OFC activity and expectation-driven behavior through temporally specific microstimulation; and (2) Elucidate the functional network dynamics between three interconnected cortical regions integral to expectation bias and error signaling—OFC, anterior insula, and anterior cingulate cortex— during expectation generation, violation detection, and behavioral adjustment processes. By simultaneously recording from multiple cortical regions with high-density electrode arrays while delivering targeted neuromodulation, this work will reveal how expectation and error signals propagate through frontal cortical networks and how this propagation can be influenced by exogenous stimulation. The basic and translational implications of this work are significant. From a basic science perspective, the proposed work will elucidate the network effects of targeted microstimulation and the causal contributions of OFC activity to motivated behavior. Translational significance lies in identifying specific neural circuit mechanisms that could be targeted to disrupt pathological expectation patterns in clinical populations. These findings will inform the development of next generation neuromodulatory therapies with enhanced spatial, temporal, and computational specificity for treating conditions characterized by persistent maladaptive expectations.

NSF Awards · FY 2026 · 2026-04

With the support of the Chemical Synthesis Program in the Chemistry Section, Dr. Martin Tomanik of New York University is developing new approaches to build complex natural products, or secondary metabolites made by plants, fungi, and microorganisms. Such small molecules provide the basis for many important medicines, agrochemicals, and materials, but many of these compounds are difficult to prepare or modify using existing chemical techniques. As a result, the potential utility of many natural products is hindered by inefficient and laborious synthetic routes. A central tenet of this work is to implement modern chemical methods that have found limited application in the synthesis of complex molecules as key strategic elements. The resulting approaches developed for natural products in this program serve as a platform to identify new bond construction tactics and expand our understanding of reactivity in complex settings. The researchers involved in this program receive rigorous synthetic training, preparing them to drive innovation across a range of chemical industries. The project also provides educational benefits by introducing high school students in the New York City area to chemistry research through outreach programs, designing lessons that integrate modern chemical methods into teaching laboratories, and supporting undergraduate success in organic chemistry through peer-to-peer learning environments. The award will support research on the development of modular synthetic platforms for the total synthesis of two architecturally complex natural products, including the talaromyolide meroterpenoids and sulfur-containing macrolactams. One major objective is the synthesis and structural reassignment of talaromyolide D and related family members using C–H functionalization and stereoretentive electrochemical sp2–sp3 cross-coupling transformations. These studies will confirm the correct molecular structure, provide access to previously unavailable congeners, and evaluate the generality of the key complexity-generating reactions across diverse terpenoid scaffolds. A second objective focuses on the synthesis of macrocyclic sulfur-containing natural products and fully synthetic analogs bearing redox-active sulfur motifs, enabling studies of their chemical activation and reactivity in complex molecular settings. Collectively, the project will establish broadly useful strategies for constructing and modifying complex molecular architectures and improve access to natural products that are otherwise difficult to obtain. 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

Mark Tuckerman and Norah Hoffmann of New York University and Nandini Ananth of Cornell University are supported by an award from the Chemical Theory, Models and Computational Methods program to create new quantum computing algorithms capable of predicting exact quantum properties of molecules under thermal conditions that exist in typical laboratory settings. The types of calculations needed for such predictions cannot be performed on classical computers, and an important overarching goal of this work is to develop new molecular simulation algorithms for quantum computers and to demonstrate their quantum advantage. The team of investigators will simultaneously incorporate thermal and quantum effects on molecular properties using an alternative formulation of quantum theory established by the physicist, Richard Feynman, based on a “sum over histories” (also known as a “path integral”), which provides a natural, yet largely unexplored, framework for performing quantum computations at finite temperature. The team will target different emerging quantum computing architectures, including spin-based, quantum resonator-based (bosonic), and topological constructions, and they will create quantum computing strategies for studying molecules in their isolated states and under the influence of structured light, which can be exploited to tune specific molecular properties. Due to the rise of quantum science as a national priority, the team will create new educational opportunities, including internships, course materials, and summer programs, designed to train next-generation scientists to work at the interface of physics, chemistry, mathematics, and computer science, in the fascinating world quantum computing. The fields of quantum information science and quantum computing have been identified as national priorities, and within these fields, quantum computational chemistry is one of the target application areas. It is, therefore, critical to anticipate the emergence of these technologies and embark on a program of developing the types of algorithms and computing strategies that will be able to leverage them and uncover and address the challenges in doing so. Among these challenges is the problem of performing quantum dynamical simulations of chemical systems at finite temperature, for which the Feynman path integral formulation of quantum mechanics is particularly well suited. Nevertheless, there have been few, if any, attempts to design quantum algorithms for path integral simulations of full molecular Hamiltonians. In this project, the team of Mark Tuckerman and Norah Hoffmann at New York University, Nandini Ananth at Cornell University, and their groups will address this challenge and develop quantum algorithms for path-integral based quantum dynamical simulations of molecular systems that are tailored to different types of quantum computing architectures. These algorithms will be designed for spin-based, bosonic, and topological quantum computing platforms and will be formulated for property prediction from quantum time correlation functions in isolated molecules as well as molecules in optical cavities subject to electromagnetic fields. Since quantum computing and quantum information science are interdisciplinary, sitting at interface of physics, chemistry, mathematics, and computer science, the team, whose expertise spans these areas, will create new education opportunities at this interface, including internships, course materials, and summer programs, designed to train next-generation scientists in the required interdisciplinary knowledge. 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 · 2026-03

Project Summary To survive, animals must learn to maximize reward and minimize the cost of time. The value of a reward is discounted according to its anticipated delay, termed temporal discounting, and the rate of this depreciation for money, food, and sex varies with hormonal fluctuations over the menstrual cycle in humans. Yet, how hormones modulate molecular and neural activity underlying temporal discounting is unclear. Closing this gap is crucial for our understanding of how endogenous hormones modulate the subjective value of rewards. Studying perturbations by hormones holds promise for revealing previously unresolved neural mechanisms that support temporal discounting. The dopamine and serotonin neuromodulatory systems are compelling candidates for the site of modulation, given their known roles in temporal discounting and deliberation between rewards with varying delays. We have previously identified alterations to the expression of critical components of the dopamine and serotonin signaling pathways over the rat reproductive cycle. I will use high-throughput behavioral training on a behavioral paradigm I developed for studying temporal discounting in rats, enabling application of powerful tools to monitor and manipulate neural circuits. I will suppress hormone activity to probe its causal role in behavior, neural activity, and molecular regulation. This proposal is organized in three Aims that encompass a K99 training phase and an R00 independent phase. In the first Aim (K99), I will optogenetically tag dopaminergic neurons and record their activity as behaving rats estimate the temporally discounted value of rewards. I will develop and apply computational models to explicitly test hypotheses about how hormones modulate discounting rate and the neural populations that represent temporally discounted future rewards. In Aim 2 (K99) and Aim 3A (R00), I will determine the molecular mechanism hormones use to regulate the expression of neuromodulatory signaling proteins, starting with the hypothesis that this regulation is at the level of protein synthesis. Finally, in Aim 3B (R00), I will study the dynamics of serotonin release with fiber photometry while rats decide between reward delays. These experiments will establish how hormones affect behavioral engagement of the dopaminergic and serotonergic systems in rats performing a rich, quantifiable decision-making behavior. The Center for Neural Science at New York University is the ideal place to address the intersection between systems-level questions about neurophysiology during behavior and molecular-level questions about protein synthesis regulation with a computational model that relates them to hormones. The strong, collaborative mentorship that I will receive from Drs. Christine Constantinople and Eric Klann will allow me to merge my exceptional experimental skillset with statistical modeling approaches, enabling the interdisciplinary approach necessary for this proposal that will become the foundation of my research program in my independent laboratory. It will have profound basic science implications for understanding hormones in the brain.

NIH Research Projects · FY 2026 · 2026-03

PROJECT SUMMARY Vocalizations and other bioacoustic signals convey an individual’s identity, location, and behavioral state and these sounds are used to guide social-acoustic interactions. While much stellar research has been done on the neural basis of vocal communication, it has largely focused on brief, stereotyped interactions between unfamiliar adults in featureless test cages. Here, we propose to study auditory communication within a cohesive social unit, the family, of a highly social species, Mongolian gerbils, in a stable, naturalistic setting. By integrating audio, video, and neural data, we aim to uncover the behavioral meaning of distinctive vocal bouts and test how cortical mechanisms support the use of these vocalizations to make social decisions. In Aim 1, we will create and disseminate general-purpose machine learning tools for studying auditory communication within family groups living in undisturbed naturalistic environments. Aim 1A will establish multi-animal 3D body pose tracking in complex spacious environments. Aim 1B will develop deep learning methods to noninvasively localize and attribute vocalizations to specific family members, focusing on the multi-second vocal bouts that dominate auditory communication. Aim 1C will integrate body pose, bioacoustic signals, and neural data with a 3D environment model using Gaussian splatting, enabling us to estimate how auditory signals are combined with line-of-sight information. In Aim 1D, we collaborate with a team at Princeton to extend and validate sound attribution methods in a different species, environment, and behavioral paradigm (mouse courtship) to establish general-purpose and open-source tools for the field. Aim 2 develops a machine learning approach to extract low- dimensional behavioral descriptors from this large data set that are used as variables in our predictive models. In Aim 2A, we extract recurring bouts of acoustically driven behavior using a mixture of expert-supervised and data-driven, unsupervised approaches to annotate continuous streams of natural behavior. Computational innovations are proposed, and we employ playback and hearing attenuation manipulations to test causal relationships between vocal bouts and behavior. In Aim 2B, we use these latent features and contextual variables to build linear Bayesian models that predict future actions, permitting us to ask how individual behavioral traits are explained by social (e.g., kinship, sex, age) and environmental factors. Aim 2C will characterize how auditory cortex activity represents vocal bouts, social and environmental factors, and subsequent behavioral decisions. Aim 3 will test the cortical network mechanisms that give rise to social and contextual modulation during auditory communication. In Aim 3A we will silence auditory or frontal cortices, as well as projections between the two, and measure how these perturbations influence behavior. In Aim 3B, we will record wirelessly from frontal cortex during sound-driven social interactions.

NIH Research Projects · FY 2026 · 2026-03

Project Abstract In order to survive nutrient deprivation and fluctuating environments, microbes adopt quiescent (non-growing) physiological states. Some quiescent states, such as bacterial spores, are remarkably resilient and can re- initiate growth after surviving centuries as a dormant spore. Compared to the exponential growth phase of bacteria, relatively little is known about quiescent states. For example, quantitative relations between growth rates and cellular physiology have been established for the exponential growth phase, but no equivalent relations have been formulated for non-growing states. To that end, the central goal of the proposed work is to uncover such relations for non-growing states through the development of data-driven, predictive biophysical models of molecular mechanisms which characterize cellular quiescence. These models will incorporate data from a variety of experimental studies, which will improve our current understanding of the interplay between certain physiological traits and molecular processes that have been observed in non-growing states. This proposal builds on our previous works on rate-limiting molecular processes in exponentially growing unicellular organisms, in which data-informed biophysical models allowed us to deduce previously unrecognized growth laws and invariant quantities. In a similar spirit, here we aim to deduce quantitative relations characterizing cellular quiescence, to provide a new framework for understanding non-growing states. Over the next five years, we will focus on three main research thrusts, all motivated by experimental observations: (i) regulation of protein synthesis and ribosome abundance in quiescent states; (ii) the influence of pH on critical metabolic reactions; and (iii) the effect of protein composition on germination rates of bacterial spores and its implications for population growth. Our biophysical models and computational analyses will be validated against data provided by experimental collaborators. This five-year plan will enable us to lay the foundations for our long- term vision to uncover principles of dormancy and deliver predictive models that can be applied in the laboratory. The proposed work will provide molecular-scale insights into the physical processes that support cellular quiescence, and will contribute to the discovery of new methods to target non-growing cells. In particular, antibiotic-tolerant persister cells present one such application.

NIH Research Projects · FY 2026 · 2026-03

Project Summary Biological accounts of reinforcement learning posit that dopamine encodes reward prediction errors (RPEs), which are multiplied by a learning rate to update state or action values. The learning rate is often assumed to be constant, but studies in humans, monkeys, rats, and mice, have found behavioral evidence for dynamic learning rates. In volatile environments, dynamic learning rates allow animals to learn faster when the world is changing, and more slowly when the world is stable. While dopamine is thought to instantiate RPEs, we recently found that dopamine release in the ventral striatum did not reflect learning rates, suggesting that dopamine-independent mechanisms determine the rate of error-driven learning. Moreover, we present strong preliminary data showing that inactivation of the orbitofrontal cortex (OFC) eliminates dynamic learning rates behaviorally, and that OFC neurons that project to the ventral striatum seem to encode the learning rate in their firing rates. In this proposal, we will determine how OFC projections to the ventral striatum dictate the rate of error-driven learning at behavioral and neural levels. This proposal will use a novel behavioral paradigm in rats, in which reward statistics vary over latent blocks of trials. We previously found strong behavioral signatures of dynamic learning rates in rats performing this task. High-throughput behavioral training will generate dozens of trained subjects for experiments in parallel, accelerating the rate of research progress. We will use optogenetics and electrophysiology to record from and manipulate OFC neurons that project to the ventral striatum, to determine if this projection pathway dictates behavioral learning rates (Aim 1). We will use electrophysiology and optogenetics to relate behavioral learning rates and activation of OFC neurons that project to the ventral striatum to trial-by-trial changes in evoked spiking in the striatum (Aim 2). We will use optical methods to measure dopamine release in the striatum and activation of OFC axon terminals, while simultaneously recording action potentials from the ventral striatum, to relate endogenous fluctuations in coincident dopamine and OFC inputs to trial-by-trial plasticity of evoked spiking (Aim 3). These experiments will test key predictions of “three-factor” plasticity rules in behaving animals. These experiments will address a major open question, which is how specific output pathways from OFC interact with downstream circuits to coordinate value-based decisions and learning. Neuromodulatory systems including dopamine are implicated in myriad neuropsychiatric disorders including schizophrenia and depression. A greater understanding of the circuit mechanisms by which they coordinate different aspects of behavior and interact holds promise for revealing novel therapeutic targets for these disorders.

NIH Research Projects · FY 2026 · 2026-03

PROJECT SUMMARY One of the brain’s most remarkable abilities is its capacity to learn and adapt continuously throughout life. This capacity relies on balancing two competing demands: keeping context- and task-specific details distinct while also integrating shared structure across experiences to enable generalization. However, we still do not fully understand how the brain concurrently manages these demands, largely due to the lack of tasks designed for studying them together. We propose that memory replay—the reactivation of brain activity patterns in the absence of overt task demand—support continual learning in the brain by reorganizing neural representations to fulfill these demands. Although past research has identified evidence of replay across animals and humans, its role in learning and behavior remain unclear. To address these gaps, we have developed a new experimental design that examines examines how the brain manages to keep task-specific details separate while also extracting common patterns during continual learning. We will also explore how memory replay enables the brain to strike this balance. Our approach tests two main hypotheses: first, that the brain forms representations that segregate context- and task- specific information while abstracting shared structure across tasks; and second, that replay supports continual learning by helping to orthogonalize and abstract task representations. In our study, participants will first learn simple, one-step transitions before planning longer action sequences across three different graphs, all while their brain activity is recorded using magnetoencephalography (MEG). This design will allow us to assess how well participants retain details specific to the individual graphs and extract a hidden, abstract structure common to all of them. We will analyze both behavior and neural patterns to understand how the brain manages these dual demands, and we will compare human behavior and neural representations with that of neural network models optimized for the same tasks. We will also use advanced MEG decoding techniques to track replay events during both rest and active task phases, examining how these events shape behavior and task representations. Complementary measures, such as eye-tracking, will help us explore how different physiological states influence replay dynamics. By combining behavioral testing, neuroimaging, and computational modeling, this study aims to provide new insights into how the brain continually adapts to changing environments. The findings will deepen our fundamental understanding of human learning and memory, and guide future efforts to enhance cognitive function in educational, clinical, and aging settings.

NSF Awards · FY 2026 · 2026-03

Modern industrial control systems are becoming increasingly interconnected through advances in digital communication, sensing, and automation. This connectivity brings substantial benefits for efficiency and reliability, but it also exposes critical infrastructure to sophisticated cyber and physical threats. Attacks on pipelines and power systems in recent years have demonstrated that a single breach in the information-technology layer can cascade rapidly into the operational-technology environment, disrupting essential services and endangering public safety. This project addresses this growing challenge by developing an integrated scientific framework for understanding how attacks propagate across these layers and by designing new methods that enable control systems to anticipate, withstand, and adapt to adverse events. Through collaboration and open educational activities, the project will help build a well-prepared workforce capable of advancing the security and resilience of future industrial systems. The technical goal of the project is to create a unified modeling, analysis, and control framework for resilient operation of industrial control systems under adversarial uncertainty. The research integrates hybrid-systems modeling, formal verification, dynamic learning, and game-theoretic decision making to capture the coupled behavior of information-technology and operational-technology components, as well as their interplay with potential attackers. The work develops analytical tools to quantify system safety and performance, introduces predictive and adaptive control mechanisms that can learn and respond in real time, and evaluates these methods through experiments on a state-of-the-art microgrid testbed. The results are expected to establish scientific foundations for resilient control, advance computational tools for assessing operational safety, and produce software prototypes and benchmarks broadly useful to researchers, industry practitioners, and critical-infrastructure operators. 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 · 2026-02

Chronic pain affects millions of Americans and remains difficult to treat without relying on opioid drugs. To develop more effective pain treatments, scientists need better tools to study pain signals in nerve cells. Recent research suggests that pain signaling occurs at receptors on the surface of neurons and can continue from inside the cell after surface receptors are internalized. Existing tools cannot directly study these internal processes. This CAREER project will develop tiny, light-responsive particles that can enter nerve cells and turn off pain-signaling receptors when illuminated with light. These particles will also emit light, allowing researchers to see where pain signals have been disrupted inside cells. This work will provide new insight into pain biology and may help identify targets for non-opioid pain therapies. The research is integrated with a strong education and outreach program designed to inspire the next generation science and engineering workforce. Overall, this project serves the national interest through biotechnology innovation, workforce development, and progress toward safe and effective health solutions. This CAREER project will develop innovative light-responsive particles to study intracellular pain signaling in sensory neurons. New light-responsive chemistries will be developed that can release bioactive molecules and generate a fluorescent reporter upon illumination. These light-responsive molecules will be combined with internal fluorescent quantum dot standards and incorporated into nanoparticles to create self-reporting probes that function inside living cells. By studying how molecular structure and nanoscale environment affect probe properties, the project will establish rules for designing efficient, self-reporting, light-activated nanoparticles. Beyond pain research, these probes will be broadly applicable tools for cell biology and biomedical research. The probes could be leveraged to identify novel therapeutic targets. In the long term, this work will enable deeper insight into chronic pain mechanisms and accelerate the development of non-opioid therapies. 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 · 2026-02

Abstract Protein synthesis is a vital biological process, important for neuronal development, synaptic plasticity, and cognitive functions such as learning and memory. In contrast, dysregulated translation is a feature of many neurodegenerative disorders, including Alzheimer’s disease (AD) and frontotemporal dementia (FTD). Pathogenic changes to the microtubule associated protein tau are thought to cause neurotoxicity and dysfunction in both AD and FTD in part by disrupting several molecular processes, including protein synthesis. However, the molecular mechanisms by which pathogenic tau disrupts protein synthesis remain elusive. In this application, we will determine how FTD-associated heterozygous mutations in tau impact protein synthesis in human neurons. We will use human induced pluripotent stem cell (iPSC)-derived neurons carrying FTD-associated tau mutations as a model. The iPSCs will be differentiated into neurons using Neurogenin-2, a master transcription factor capable of inducing differentiation into excitatory neurons in under two weeks. Using this platform, in the first aim we will determine the impact of FTD-associated tau mutations on translation elongation rates and will perform ribosome profiling to determine the translatome and translational efficiency associated with the FTD- associated tau mutations. In the second aim, we will determine whether the tau mutations alter tau- ribosome interactions and if they cause ribosome collisions. These studies will provide insight concerning the mechanisms by which FTD-associated mutations in tau alters protein synthesis, as well as the biology and subsequent pathobiology of tau in tauopathies such AD and FTD.

NSF Awards · FY 2026 · 2026-01

Modern scientific computing increasingly depends on high-performance, reusable software components known as libraries, which are often written in low-level languages such as C and C++. While these languages offer speed and flexibility, they are also prone to memory corruption, crashes, and silent data errors. Such failures can compromise critical scientific workflows, leading to inaccurate or unreliable research outcomes and wasted computational resources. GRISL (General-purpose Rigorous Isolation for Science Libraries) strengthens the robustness and security of cyberinfrastructure by isolating and hardening scientific libraries through a lightweight, userspace-based containment approach. GRISL enables scientific researchers, many of whom are not systems experts, to continue using legacy libraries without modifying their code, while gaining protection against memory bugs and data integrity issues. By making these libraries safer by default and compatible with widely used scientific platforms, GRISL improves the reliability of artificial intelligence and machine learning applications, high-performance computing systems, and domain-specific simulations in fields such as physics, biology, and climate science. Technically, GRISL introduces a novel form of protection to wrap and monitor unsafe libraries with minimal performance impact (~1%). Its architecture enforces advanced runtime safety checks that are helpful not only for traditional scientific libraries but also for Large Language Models (like ChatGPT). GRISL also protects different libraries from each other, by providing a safe way to perform inter-library communications. These innovations allow researchers to run hardened versions of popular scientific libraries with minimal overhead, and validate their functional correctness using advanced testing techniques. The project collaborates closely with national testbeds such as CloudLab and Chameleon. By preventing crashes and silent corruption, GRISL improves scientific reproducibility and resilience across computational disciplines, and will be broadly disseminated through open-source releases, educational outreach, and integration with scientific ecosystems. 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

A graph is a collection of vertices (points or objects), and a collection of edges (links or lines), that connect pairs of vertices. Graphs are a central and an extensively studied type of mathematical object, and they are commonly used to model various problems in many different real world scenarios and applications. For example, it is natural to model a road network in a city, or a computer network, or friendship relationships in a social network as a graph. There are countless other scenarios where a problem one needs to solve, or an object one desires to study, can be naturally abstracted by a graph. As a consequence, the design of efficient algorithms for central graph problems is fundamental to computer science and beyond, and has a significant impact on many aspects of computation. As the amount of data that applications need to deal with grows, it is increasingly important to ensure that such algorithms are very fast. In this project, the investigators will study several central graph problems, such as Maximum Matching, Maximum Flow, and Shortest Paths, in two basic settings. The first is the standard model where the input graph is known in advance, and the goal is to design a fast algorithm for the problem, with running time not significantly higher than the time required to read the input, which is close to the fastest possible running time. The second is the model of dynamic algorithms, where the graph changes over time (for example, consider a road network, where the computation has to account for roads becoming more or less congested with traffic), and the goal is to quickly support queries about the graph, such as, for example, computing a short path between two given vertices. This project is organized along four main interconnected thrusts. The first thrust focuses on the design of algorithms for dynamic All-Pairs Shortest Paths (APSP), that can withstand an adaptive adversary, and that significantly improve upon the currently known tradeoffs between the approximation quality and the running time, in both directed and undirected graphs. Algorithms for APSP and its variants are often used in combination with the Multiplicative Weights Update framework to efficiently solve various flow and cut problems in graphs, and thus provide a valuable and powerful algorithmic toolkit. The second thrust is directed towards improving and extending known expander-related tools that are often used in the design of fast algorithms for various graph problems. Expanders are playing an increasingly central role in graph algorithms, and these tools can serve as building blocks for many other graph problems. The third thrust focuses on the Maximum Matching problem. Using techniques inspired by algorithms for dynamic shortest path in directed graphs, the goal of this part of the project is to develop fast combinatorial algorithms for both the bipartite and the general version of the problem. The final thrust focuses on designing improved algorithms for maintaining near-optimal matchings in dynamic graphs, building on insights and algorithms developed for the second and the third thrusts. 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 · 2025-12

We all are regularly asked to make decisions in the face of uncertainty: we decide how best to drive to work or how and at what prices to buy and sell things (everything from houses to groceries). And our computers make such choices as well to give good performance to their users: which order to process tasks q to maximize the throughput or minimize user delay, or which pages to keep in fast memory (because they will be accessed again soon) and which others moved to slower memory. These questions do not fall in the classical ``one-shot'' framework of computation, where an algorithm reads in the input, processes it and produces its output. Instead, they fall in the area of sequential decision-making, and specifically of online algorithms, which provides a framework in which we can formalize such questions and also reason about their solutions. For such problems, there is a natural tension between two competing desires: (a) to make decisions that maximize the instantaneous gratification but may be poor in the long run, or (b) to wait and maneuver into a good position to make gains in the future. This project aims to develop general-purpose algorithms that find optimal ways of hedging between these two extremes for a broad class of optimization problems in online optimization. As an example, consider the k-server problem, which is a central problem in this area. In this, one controls k servers which move between some locations in a metric space. A sequence of requests arrives over time, where each request specifies a location, and one must move one of the servers from its current position to this requested location. The goal is to minimize the total server movement. Given a request, which server should be moved? As mentioned above, there is a tradeoff between being greedy and moving the servers closest to the requested location and moving more distant servers to be in a better position for future requests. Developing good algorithms for this and related problems has been a major challenge in the area. This project will investigate three ways of addressing the challenge: (a) To develop general principles for designing broadly-applicable algorithms for randomized settings, and in particular, to extend the classical work function algorithm (which is near-optimal for the deterministic setting) to randomized settings; (b) To get extendable, robust convex relaxations for k-server and its generalizations, and to use these relaxations to obtain good algorithms; (c) To develop a broader framework for typical instances, instead of focusing on the worst-case instances. Our work will draw on ideas from linear and convex optimization, stochastic optimization and optimal stopping theory, and online learning in order to deepen our understanding of these and other central questions in optimization under uncertainty. 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 recent technological advancements and shifts in work patterns, traffic congestion continues to burden U.S. cities, resulting in the loss of millions of dollars due to wasted fuel and lost productivity. In congested urban areas, these issues are exacerbated by the time and stress associated with searching for parking. The standard strategy toward travel-demand management to tackle traffic congestion is road and parking pricing, but the real-world implementation of pricing approaches has proven to be difficult due to concerns about their fairness. This project proposes an alternative in the form of integrated parking-management and ridesharing systems that reduce traffic congestion through non-pricing mechanisms. In addition to alleviating congestion, the proposed systems aim to expand access to opportunities through collaborative ridesharing solutions that are deployable even in auto-dependent communities with limited transit options. This project aims to develop a data-driven travel-demand management framework for the deployment of an integrated parking-management and ridesharing cyber-physical system (CPS). The CPS will enable the implementation of integrated parking-management and ridesharing schemes to mitigate congestion and enhance mobility by jointly coordinating parking access and collaborative peer-to-peer carpooling programs. The proposed framework advances research on peer-to-peer ridesharing and travel-demand management through optimization models that jointly consider hybrid travel-incentive structures, measures of personal affinity between travelers, and user-specific valuations of these measures in the ride- and parking-matching process. The framework envisions a CPS architecture that dynamically updates users’ affinity measures and accounts for trade-offs they make between travel costs and trust considerations. Beyond new policies, models, and heuristics, the innovation of this project lies in the proposed CPS, which combines traveler digital twins, machine-learning-based predictions of parking-facility states, and internet-of-things technologies to operationalize the proposed travel-demand management schemes. Lastly, the project will establish a living laboratory to evaluate the performance of the integrated parking and ridesharing CPS under real-world conditions. 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

Accurate interpretation of hyperspectral data depends on the availability of reference spectra: measurements of known materials compiled into spectral libraries. Such libraries support both direct classification and machine learning applications. When combined with on-site hyperspectral imaging, they have proven effective across a variety of domains including heritage conservation, homeland security, hydrology, and geology. Urban conditions, however, present unique challenges to spectral data collection. In this context, urban materials refer to the components of the built condition, including both manufactured materials (e.g., asphalt, concrete, paint) and naturally occurring materials that have been anthropogenically modified for urban use (e.g., cut stone). Although hyperspectral data have been utilized in select urban planning tasks, the broader potential of hyperspectral imaging for material identification remains underutilized. This project will develop and extend community involvement in hyperspectral remote sensing technology to analyze and study urban landscapes. This will be paired with open metadata standards, modular processing toolkits, and automated archival workflows that prioritize FAIR principles. HS-SPECTRA (Hyperspectral Standardizing and Sharing Possibilities for Urban Conditions through Toolkits, Resources, and Archiving) addresses fundamental challenges in hyperspectral library design by: 1) Developing a metadata architecture tailored to longitudinal urban field campaigns; 2) Incorporating auxiliary sensors to contextualize spectral variability; 3) Implementing a flexible versioning and querying model that reflects the dynamic nature of repeated, real-world observations; and 4) Enabling interoperability across platforms through spectral resampling and standardization pipelines. By embracing the temporal complexity of real urban conditions and focusing on reproducible, extensible data infrastructures, HS-SPECTRA will generate a uniquely valuable dataset for cross-instrument, cross-temporal analysis. The resulting protocols and open-source tools will significantly advance methodological rigor and accessibility in urban planning, Earth system science, computer vision, and urban monitoring fields. This award by the Office of Advanced Cyberinfrastructure within the Directorate for Computer and Information Science and Engineering is jointly supported by the Directorate for Engineering. 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 will build the first multiscale AI ocean emulator, spanning from submesoscales to large-scale ocean flows, for long-term ocean variability. The emulator will be used to investigate the complex nonlinear dynamics of the ocean circulation and to bridge the gap between understanding and simulating ocean variability across many time and space scales. This project will also advance the development of physics-informed, autoregressive neural networks capable of learning from heterogeneous, multi-resolution datasets. It addresses core challenges in scientific machine learning, including stability and interpretability in complex systems, and introduces methodologies that can extend to other multiscale problems. Understanding the interactions between oceanic processes across scales is essential for advancing our knowledge of ocean circulation, heat and momentum transport, and their role in shaping long-term variability. The primary objective is to investigate the spatio-temporal interactions between submesoscales, mesoscales, and large-scale flows on regional and global scales. This will be achieved by building a three-dimensional multiscale AI emulator, using deep neural networks, trained on a suite of numerical and observational datasets at different spatio-temporal resolutions. The questions to be addressed include: How to construct a physically-based 3D AI ocean emulator from heterogeneous datasets? How to evaluate a multiscale emulator from sparse and imperfect datasets? What fraction of ocean submesoscale and mesoscale physics drives momentum, energy, and heat transport at large scales? What is the role of multiscale processes on local ocean variability, such as marine heat waves and sea level extremes? 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.