Columbia University

NSF Awards · FY 2025 · 2025-08

With the support of the Chemical Catalysis Program in the Division of Chemistry, Professor Tomislav Rovis of Columbia University is studying the development of new chemical reactions to solve challenging problems that pharmaceutical scientists encounter in collaboration with Bristol Myers Squibb (BMS) researchers. New therapeutics contain increasingly complex chemical structures that necessitate innovative methods to construct them. Chemists doing drug discovery research must either make these complex molecules through longer routes, or choose to make alternate, less desirable structures. Prof. Rovis and the BMS team are using catalysts activated by low energy orange light to drive chemical reactions that promise to be more tolerant of reactive functionality and complex structures, facilitating the assembly of molecules having desirable biological properties. This project is also facilitating the scientific interactions and exchange of ideas between BMS scientists and Columbia graduate students, with regular research updates and quarterly reciprocal visits strengthening relationships while exposing junior scientists to industrial science. These activities are aiding chemists everywhere to make more complex therapeutics of tomorrow. Reactions that forge alkyl-aryl carbon-carbon bonds are of high importance in modern drug design to prepare compounds with improved physicochemical properties through shorter synthetic routes. Prof. Rovis and his research group are collaborating with BMS to design catalytic systems that use visible light to drive these reactions. By harnessing novel catalysts, lower energy orange light (560-600 nm) is being used to couple two distinct components to make a new carbon-carbon bond under mild and potentially more tolerant reaction conditions. In an orthogonal approach, the team is using cheap and widely available copper catalysts to make an emerging functional group increasingly found in bioactive compounds by facilitating a challenging reductive elimination. This transformation is also driven by visible light, translating the energy in a photon to accelerate a desired bond construction. Lastly, a machine learning platform is being developed and exploited to predict optimal catalytic conditions for carbon-carbon bond formation in photocatalytic systems using acetyl acetonate ligands. 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 · 2025-08

Predictive modeling can accelerate the design of fuel-efficient, fuel-flexible engines important to a robust energy future, industry competitiveness, and a resilient national defense. To meet these needs, models must accurately represent the key reactions that occur during combustion of the fuel inside the engine. Such models would be especially useful for ammonia, a fuel of significant recent interest with unusual combustion characteristics. However, substantial gaps remain in understanding ammonia combustion, particularly at the operating conditions of advanced engines. The goal of this project is to create and validate a model for ammonia combustion that is specifically tailored to make accurate predictions at desired engine conditions. A novel approach, which leverages modern computational chemistry, data science, and optimal experimentation, will be used to identify, confirm, and characterize the chemistry important to desired applications. The resulting model will provide a valuable predictive tool for designing ammonia-powered engines while also establishing the foundation for other nitrogen chemistry models. Likewise, the overall methodology will aid future investigations of many other chemically reacting gases. This project also provides student research opportunities in an interdisciplinary, collaborative setting and partners with industry to enable research to lead to better engine designs immediately. The technical objective is to create a multiscale, physics-based, data-driven model for ammonia combustion by optimally selecting, creating, and exploiting data, including data for new chemical pathways. The approach integrates (1) automated theoretical calculations to identify and characterize previously undiscovered chemistry impacting engine predictions, (2) targeted experimental measurements to validate new chemical pathways of importance and optimally inform engine predictions, and (3) uncertainty-quantified modeling based on theoretical and experimental data to create models that are physically sound and accurate. Ultimately, the research will address key knowledge gaps in ammonia combustion and produce the first ammonia combustion model whose reaction sets and data-informed parameters are specifically tailored to meet the needs of engine design. The new data, models, and multiscale data-driven approach will also enable better predictive understanding of many other complex reacting systems, including energetic materials/propellants, planetary atmospheres, and hypersonics. 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 · 2025-08

The research in this project investigates how scientists’ decisions – ranging from the type of AI tools and models to use, to how to train models and label behaviors – can shape what neuroscientists come to know about animal minds. Drawing on philosophy of science, neuroscience, and science and technology studies, the project analyzes how decisions about data, software design, and interpretation affect scientific outcomes. The study’s overarching goal is to improve how behavioral tracking tools are used, thereby making research more thoughtful and effective. The findings of this project will be of interest to scientists, educators, designers and users of AI. This project conducts a comparative assessment of several widely used AI-based behavioral tracking tools to investigate how different algorithmic approaches shape the study of behavior. The tools are built on distinct algorithmic foundations, ranging from supervised learning techniques that track visible features, such as body pose, to unsupervised models that infer behavioral patterns over time. Because many of these tools rely on machine learning, the project contributes to a better understanding of how artificial intelligence is now shaping scientific inquiry. The research examines how differences in these technical methods influence how behavior is organized, categorized, and interpreted. It addresses how behavioral knowledge is developed and defined, in turn offering pathways for developing more transparent, flexible, and representative research across species and settings. 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 · 2025-08

This project aims to make the Internet more reliable by giving operators more control over how data reaches their network from other parts of the Internet. Today, network operators rely on trial and error, and often settle on subpar outcomes, because they do not have direct control over or visibility into which geographic locations and other networks the data will traverse to reach their network. This project aims to change that by creating a system where operators can simply specify their goals, such as improving performance or avoiding particular countries, and then the system figures out how to make that happen. This collaborative project brings together investigators from Columbia University, Northeastern University, and Federal University of Minas Gerais (UFMG) in Brazil. The project envisions a system that allows a network operator to describe the policy for how to set preferences for possible ingress routes (which the project calls “intents”), then automatically configures routing announcements to achieve the most preferred (feasible) outcome. To realize this vision, the project will address the following research questions: 1) What intents are desired, and how can it be made easy for operators to express them? 2) How can a system predict the routes and traffic engineering metrics that will result from an announcement? 3) How can a system automatically learn which configurations are possible and what their semantics are? 4) What are efficient ways to search through large numbers of configurations spanning multiple networks to satisfy general intent? The project will enable new ways for Internet providers and cloud services to improve the reliability and performance of Internet services on which society increasingly relies. For example, our project can help Internet providers maintain service during and after natural disasters, as well as identify and block Internet attacks. It will help improve performance for a wide range of services that include online educational technologies, telemedicine, remote work, and/or various forms of e-commerce and entertainment. The research outcomes will also serve as a foundation for future academic and industrial innovation in Internet routing. Project updates and outcomes will be published at https://ingress-routing.ee.columbia.edu/, with the plan to maintain the site for at least three years beyond the award period. 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 · 2025-08

Coastal regions are vulnerable to flooding from rivers and rising seas, increasing storm strength, and destruction of ecologically-fragile areas. River deltas are especially impacted by the balance between increasing water levels from sea-level rise and tides, and land surface elevation changes. Bangladesh’s Ganges-Brahmaputra Delta (GBD), the world’s largest delta, is a particularly excellent place to investigate this problem. The land is sinking (subsiding), worsening the impact of sea-level rise, but the rivers supply ample sediment to elevate the land. However, there is a mismatch in the distribution of sediment and land subsidence; some areas are maintained by sediments, while others are at serious risk of land loss. This project will combine local, on-the-ground measurements of elevation change with broad satellite observations, and develop a comprehensive numerical model of elevation change. The numerical model will enable synthesis of all measurements and incorporate shallow processes that are missing from most models. Results will contribute to Bangladesh’s coastal planning through established collaboration with government agencies, academic institutions, and non-governmental organizations. This project will support 2 postdocs and 3 graduate students in the U.S. as well as build capacity for students and faculty in Bangladesh. U.S. undergraduate students will participate in the proposed research through internship programs and a capstone course that includes a Spring Break field trip to Bangladesh. The model will have great applicability for use in coastal areas prone to flood risk, especially lowland deltas worldwide including the Mississippi Delta. Unraveling the intersecting processes that contribute to vertical land-surface dynamics is critical for forecasting sustainability of lowland deltas into the future. This project will employ multidisciplinary research that integrates an existing delta-wide network of sediment cores and geospatial instruments with broad-scale, multi-sensor satellite remote-sensing analyses, producing novel high-resolution maps of decadal surface-elevation change, topography, and land-use across the coastal zone. A state-of-the-art poroelastic model will be developed, validated, and applied to coastal Bangladesh. The team hypothesizes that at any given site on the delta, surface-elevation change reflects the vertical integration of sedimentation, near-surface soil consolidation, subsurface compaction of Holocene sediment, and deep tectonic/isostatic response of the lithosphere. Across the delta, surface-elevation change reflects how modern land use restricts surface sedimentation and accelerates consolidation, and how ancient river dynamics constructed the alluvial architecture of compacting Holocene sediments. These hypotheses will be tested with a process-based, holistic understanding of vertical land-surface dynamics, and will guide coastal hazard mitigation and sustainability efforts on the GBD and other deltas that face similar environmental and anthropogenic stressors (e.g., Mississippi and Sacramento-San Joaquin river deltas). 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 · 2025-08

This award funds a research project that examines the long-run relationship between material resource usage and economic growth in the United States. It challenges the conventional view that rising living standards inevitably demand increasing material consumption. By integrating theoretical insights with data analysis, the project investigates the mechanism and the impacts of material efficiency in aggregate production. Its theoretical and empirical examination of material efficiency from the macroeconomic perspective represents a novel and major contribution to literature. One of the main contributions of this award is that it constructs a real, chain-weighted index of over 150 raw material inputs—spanning energy, major and minor metals, biological commodities, and construction materials— to document material consumption since the early 1970s in the United States. This project is important for understanding how innovations in material-saving technologies affect macroeconomy, public welfare, and development. Material-use efficiency underpins many advanced manufacturing technologies and supports translation research. The research on material usage carries profound implications for economic growth and national security. The research work could advance scientific understanding of resource efficiency, support workforce development by creating open-access data and tools for researchers and educators, and inform decision makers to promote a resilient, prosperous US economy. Ultimately, the research findings could help the United States develop more efficient production and material-use strategies, improve citizens’ wellbeing, and strengthen its dominant position in the global economy. This award funds a research project with two integrated phases. In the first phase, the project develops a harmonized historical database of material inputs into the U.S. economy, drawing on quantity and price series from the U.S. Geological Survey, the Energy Information Administration, the Bureau of Economic Analysis, and archival sources. Real, chain-weighted input indices are calculated and mapped through harmonized, expanded input-output tables into consumption and investment sectors. Preliminary analysis of these data reveals two distinct eras: rapid material-intensive growth through mid-century, followed by persistent “degrowth” of material throughput since 1970. In the second phase, the project introduces a structural macroeconomic model that combines non-homothetic constant-elasticity-of-substitution (CES) preferences with directed technical change across material and labor inputs. The model endogenizes innovative choices by representing firm-level production efficiencies as Pareto-distributed draws and allows expenditure shares to vary flexibly with income via sector-specific income elasticities estimated from the Consumer Expenditure Survey. Sectoral elasticities of substitution and technology parameters are identified using Bureau of Labor Statistics price data, industry markups, and productivity statistics. Counterfactual simulations decompose the observed decline in material throughput into five channels: classical factor substitution, material-saving technical change, shifts in sectoral expenditure patterns toward less material-intensive services, changes in trade patterns that offshore embodied resource use, and a residual component capturing unobserved factors. The anticipated deliverables include a publicly accessible database of historical material-input series; expanded, time-consistent input-output tables; estimates of key preference and technology parameters; and a rigorously identified decomposition of U.S. dematerialization. The results could inform decision makers for both national defense and economic well-being by assessing resilience under resource scarcity, and support long-run prosperity by identifying leverage points for resource-efficient technological innovation. 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 · 2025-08

This project helps explain why volcanoes in Earth’s subduction zones erupt explosively. Subduction zones are major fault regions where the crust and upper mantle, are pulled into the mantle below. Volcanoes at subduction zones have high amounts of silica and volatiles, which make them prone to explosive eruptions. The ‘Ring of Fire’ around the Pacific rim, includes active volcanoes in Alaska, the Cascades, and Mexico that are linked to subduction. Eruptions of subduction zone volcanoes can devastate communities, halt air travel, cause crop failures, famine, social unrest, and lead to large-scale human migration. Scientists aim to understand what causes volcanic eruptions. This knowledge helps improve predictions and reduce destructive impacts. This project investigates the origin of silicic, volatile-rich melts in young Mexican volcanoes. Do they form deep within the Earth’s mantle, or do they develop in the crust above the mantle? Researchers will use the compositions of minerals from silicic magmas to address these questions. These minerals, called ‘olivines’ and ‘Cr-spinels,’ are less than a millimeter in size. The temperature and composition of the magmas will also be determined using high-precision analytical methods. Modeling this information will clarify how the magmas form and how they contribute to explosive eruptions. This project will offer many learning opportunities for high school and undergraduate students. A high school teacher will design classroom materials to teach high school students about volcanic hazards. Students will also learn how tiny crystals can inform scientists about magmas formed deep beneath Earth’s surface. Hands-on, research projects will provide undergraduate students with valuable research experiences. An interactive exhibit based on this research will be developed for the annual Lamont Open House event. Olivines with Fo>88 are recognized for their potential to contain unique information on mantle lithology, composition, and melt processing, which cannot be obtained from bulk rock studies. This potential is especially valuable for arc magmas, where Cr-spinel-bearing forsteritic olivine phenocrysts or antecrysts are often the only remaining clues of mafic melt input. However, there is reasonable doubt whether olivine trace element patterns can withstand ‘transcrustal overprinting’ caused by fractional crystallization, melt mixing, and diffusion during melt ascent through the crustal basement. The Transmexican Volcanic Belt (TMVB), where high-Mg# and high-MgO magmas erupt through thick continental crust with minimal contamination, provides an ideal setting for testing whether olivines can preserve primary melt signatures and what this implies for mantle wedge processes. Over many years of fieldwork and collection of bulk rock and crystal-scale compositional data, we have assembled a carefully selected, well-characterized set of Cr-spinel-bearing olivines from 36 samples that represent the broader spectrum of high-Mg# magmas erupting at both the arc front and the rear-arc of the Transmexican Volcanic Belt. The He-O-Os systematics of these olivines constrain their crystallization in (near-) primary mantle melts. The main aim of this project is to select a subset of already mounted and pre-characterized olivine + Cr-spinels for new high-quality laser-ablation ICP-MS analyses, targeting elements such as Li, Na, Mg, Al, Si, P, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Y, Zr, and Yb. High-Fo>88 olivines will be tested for their ability to retain the trace element spectrum of their (near-) primary equilibrium melts by correlating olivine composition with bulk rock diversity, the source-sensitive Cr# of Cr-spinel inclusions, and by comparing olivines from intraplate versus mid-ocean ridge settings. Selected TMVB olivines with Fo<88~80 will be used to evaluate whether primary olivine signatures survive transcrustal processes. Since the Trans-Mexican Volcanic Belt has an end-member character within the global spectrum of arcs, the results from this study are relevant not only for arc olivines worldwide but also for olivine research in intraplate and mid-ocean ridge 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 2025 · 2025-08

Algorithmic and machine learning tools are increasingly central to high-stakes decision-making across domains such as healthcare, finance, and digital services. These tools depend on data that are often incomplete, unbalanced, or poorly aligned with the populations they impact. While existing methods typically treat datasets as externally fixed and data imbalances as externally-determined, this project views data as the outcome of a strategic production process that is shaped by the incentives of individuals, platforms, and institutions. By leveraging these incentives, the project aims to design data ecosystems that yield more representative and reliable datasets, better reflecting the needs of all Americans. The research will develop new theoretical foundations and algorithmic methods that improve data quality through incentives rather than constraints. It will also support education on data economics in the context of AI and foster early-career development through workshops and mentoring activities. The project integrates ideas from computer science, economics, and operations research to analyze incentive-driven data production and exchange. It investigates three key dimensions: (1) how individual data providers can shape outcomes through strategic data sharing; (2) how platforms can build robust data markets by aligning incentives between producers and consumers; and, (3) how large-scale effects such as platform competition and user interactions influence data quality and decision performance. These efforts will advance the scientific understanding of dynamic data environments and inform the design of more effective data-driven 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 · 2025-08

Serpentine minerals are important to Earth's natural water cycle, especially deep underground. They form weak fault zones along tectonic plate boundaries. They carry water into the planet at subduction zones. And they play a role in forming natural hydrogen, which is a potential energy source, “food” for subsurface microbial communities and a potential step in the origin of life. But their complex structure and delicate layers have made it difficult for scientists to study them in detail. This project uses powerful computer simulations, including advanced artificial intelligence tools, to better understand how serpentines behave under high pressure and temperature. These simulations will help explain how water moves through the Earth's interior, how serpentines break down and possibly trigger earthquakes, and how they affect the signals we detect in seismic studies. This research will also develop new tools to explore how water interacts with minerals deep inside the planet. Serpentine minerals are key players in Earth’s geological water cycle, facilitating water uptake and release, contributing to mantle hydration, and influencing the rheology of subduction zones and other tectonic plate boundaries. Moreover, serpentine formation is often accompanied by oxidation of ferrous iron, together with formation of highly reduced fluids and H2 gas, in turn facilitating abiotic organic synthesis, sustaining a subsurface microbial ecosystem, and potentially providing a source of natural, “green” energy. However, the structural complexity and weakly bonded layers in serpentines have long posed challenges for both experimental and computational studies of such processes. This research will leverage cutting-edge yet proven materials simulation techniques, including advanced exchange-correlation functionals and machine learning potentials (MLPs) for large-scale molecular dynamics (MD) simulations, to investigate the phase relations and thermoelastic properties of serpentines. These are critical for understanding their formation and dehydration in subduction zones and their potential role in earthquake generation. This project will also yield seismological parameters of serpentinized mantle, helping to constrain water fluxes at subduction zones and elucidate Earth’s deep water cycle. Expanding MLP development for the more general MgO-SiO₂-H₂O (MSH) system will enable exploration of diverse hydrous environments and water–mineral interface behavior throughout the mantle and near the surface, where serpentine is linked to processes of technological interest such as carbon mineralization, hydrogen generation, and production of Ni-rich and Co-bearing alloys critical for batteries. 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 · 2025-08

This project develops powerful new methods for integrating diverse and evolving datasets, a critical challenge in modern science and technology. In an increasingly data-driven world, information is often collected from many sources, at different times, and in various formats, making it difficult to analyze as a whole. This research will create flexible and reliable tools that can automatically integrate complex information. The tools will be applied to assess the safety of autonomous vehicles by combining limited test data from a new city with vast amounts of driving data from other regions. The project will also develop open-source software for all researchers and create educational materials to train the next generation of scientists and engineers. This research will establish dependable methodologies and solid theoretical foundations for data integration through three interconnected thrusts. First, the investigator will develop transfer learning methods for integrating samples from multiple sources to enable knowledge distillation and transfer across heterogeneous datasets. Second, the research will address the challenge of data streams that evolve over time by creating techniques that adapt to temporal distribution shifts. Third, the investigator will create principled approaches for learning latent structures by integrating multiple data views of the same subjects, such as combining social network information with individual user profiles. A key intellectual contribution of this work is the development of practical procedures that automatically adapt to unknown data heterogeneity with theoretical guarantees under minimal assumptions, overcoming a major limitation of current methods. The project will deliver innovative analytical tools, new theoretical insights into data integration, and open-source software packages to benefit the broader scientific community. 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 · 2025-08

Non-technical description: The semiconductor technologies that pervade modern society work by transporting energy carriers like electrons and excitons from source to target. Carrier scattering with impurities and vibrations is the primary source of resistance in semiconductors, limiting the speed and efficiency of all electronic devices ranging from photovoltaics to computer chips. Rising needs for energy efficiency and computing power provide a strong impetus for identifying materials and processes that minimize these scattering losses. This CAREER project leverages novel light-matter interactions to minimize or completely suppress scattering at ambient conditions in two-dimensional semiconductors of high current interest, towards high-speed, nearly resistance-free transport of energy and information. For the educational component of this project, the research team develops affordable but high-performance optical microscopes capable of detecting single molecules. The microscopes are used in a hands-on microscope building workshop for high school students, and in a new undergraduate laboratory on weighing single molecules with light. Technical description: When light couples strongly to collective dipole excitations, light and matter excitations are renormalized into part-light part-matter eigenstates known as polaritons. Polaritons inherit the strong nonlinear interactions of matter and the long-range coherence of light. Although seemingly ideal for energy and information transport, polaritons are short-lived and challenging to extract. This CAREER project focuses on leveraging polaritons to modify the properties of adjacent non-polaritonic charges and excitons, even in the absence of light. This approach circumvents the short lifetimes and poor extraction efficiency of polaritons. The first aim of the project seeks to suppress charge carrier–defect scattering in disordered van der Waals semiconductors through strong coupling between light and electronic excitations. The second aim harnesses phonon-polaritons in polar dielectrics to accelerate exciton and charge transport in two-dimensional semiconductors across a van der Waals gap. In both cases, the research team uses non-contact ultrafast optical microscopy to directly visualize the motion of charges, excitons and polaritons moving over 10 nanometers–50 microns and 30 femtoseconds–10 microseconds, providing mechanistic insight into polariton-assisted carrier transport. These measurements are complemented by photocurrent and traditional contact measurements to verify compatibility with standard semiconductor hardware for applications in microelectronics. 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 · 2025-07

This project includes a multi-faceted analysis of marine carbonyl sulfide emissions to be coordinated by a collaborative team from institutions in the US, Germany, and Israel. Carbonyl sulfide is a trace gas capable of providing insight into the global carbon cycle. Mass balance estimates from isotopes and atmospheric inversions both suggest the missing sources of carbonyl sulfide are tied to marine fluxes. This effort will significantly increase understanding of the marine source of carbonyl sulfide (OCS) by conducting a series of coordinated experiments that combine: (1) direct marine flux measurements of OCS; (2) dissolved measurements of OCS and its isotopologues and precursor gasses; and (3) data assimilation and modeling. The project includes an extended field campaign to continuously measure the direct fluxes of OCS. The team will collect data using an air-sea interaction tower on the US Atlantic seaboard (near Martha’s Vineyard, MA) and in Bolkins Eck (Bering Sea), as well as using shipboard measurements to quantify fluxes and resolve sources (via sulfur isotopes) of OCS from these coastal sites. The project includes training for early-career scientists and graduate students from a team of experienced scientists. This work is supported by the Atmospheric Chemistry and the Chemical Oceanography Programs. 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 · 2025-07

This project is developing an inexpensive and accurate point-of-use test for per- and polyfluoroalkyl substances (PFAS) at drinking water facilities. The very nature of the strong carbon-fluorine bond in these molecules makes them persist in the environment and the human body. There is growing evidence that forever chemicals cause cancer and other diseases, and it is expected that stricter guidelines on the levels of PFASs allowed in drinking water will be implemented in the future. Current methods for testing for PFASs in drinking water must be carried out off-site, rely on expensive specialized equipment, and have a turn-around time of days to months. The TrueBlue sensing devices developed here will combine the ability of biological systems to recognize virtually any pollutant with high sensitivity combined with readout electronics in a small disposable “dipstick” format; sensing can occur within 1-2 hours. One of the research teams of this collaborative effort will design these hybrid bioelectronic devices to detect PFAS compounds using the latest advances in artificial intelligence. This team’s efforts are complemented by a world-leading applied research team that has been working with the New York City (NYC) Department of Environmental Protection for nearly three decades to advance water resource recovery in NYC, one of the largest metropolitan regions in the world. The intellectual merit of this project comes in delivering a transformative technology for detecting PFASs and other pollutants in water and advancing the design and application of hybrid bioelectronic devices using artificial intelligence. The broader impacts span from improving human health by minimizing exposure to PFAS compounds in the environment to rich training opportunities for future generations of STEM researchers. This project seeks the convergence of the fields of synthetic biology, bioelectronics, machine learning, and water resources engineering to provide point-of-use sensors for assessing per- and polyfluoroalkyl substances (PFAS) in drinking water. Because of the unique strong carbon-fluorine bond, PFAS, or "forever chemicals", persist in the environment and the human body, causing cancer, liver damage, immune system disruption, and developmental issues in children. Current methods for testing for PFASs in drinking water must be carried out off-site, rely on expensive specialized equipment, and have a turn-around time of days to months. In contrast, the TrueBlue biohybrid devices developed here can be utilized on-site and will use equipment that is 100-fold less expensive. This transformation will be achieved by building an entirely new class of biohybrid device that co-opts the specificity of a yeast G-protein coupled receptor (GPCR), the sensitivity and remote sensing capabilities of complementary metal-oxide-semiconductor (CMOS) read-out devices, and state-of-the-art machine learning algorithms to de novo design the GPCRs to recognize key PFAS compounds. The project includes manufacturing scale-up to provide a minimal viable product for commercialization and drives the product to start-up company creation, including the development of a go-to-market plan, regulatory plan, and commercialization plan. 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 · 2025-07

This research aims to advance our understanding of the dynamic interaction between mantle flow, crustal deformation, and topography beneath Antarctica by developing a high-resolution geodynamic model using the mantle convection code ASPECT. The model simulates mantle flow and crustal dynamics from the time of the Antarctic Ice Sheet's inception (34 million years ago) to the present. Key objectives include addressing the evolution and uplift of the Marie Byrd Land hotspot, investigating the interplay of mantle processes and tectonic forces during rift evolution in the West Antarctic Rift System, and evaluating how buoyancy-driven mantle flow interacts with glacial isostatic adjustment. The model will provide critical boundary conditions, including heat flow and topographic changes, for improved ice sheet and glacier modeling. This work will also contribute to understanding the interactions between the solid Earth and the climate system. Antarctica’s vast ice sheet sits on a continent shaped by deep Earth processes that have influenced its landscape over millions of years. This project uses advanced computer simulations to study how slow-moving mantle rock beneath Antarctica affects the ground beneath the ice and interacts with tectonic forces and the climate system. By combining seismic imaging data with cutting-edge modeling techniques, researchers will recreate the history of these processes, starting from when Antarctica first became covered in ice 34 million years ago. The work will help answer questions about why some parts of the continent are rising, how tectonic forces shaped its rift systems, and how the Earth responds to changes in ice sheet size. Results will improve predictions of future sea level rise by providing better estimates of how the ground beneath the ice changes. The project also aims to make scientific tools freely available and increase diversity in polar science by supporting early-career scientists. This research will not only uncover Antarctica’s past but also contribute to understanding its future in a warming world. Broader dissemination will occur through collaborations with the Scientific Committee for Antarctic Research (SCAR) INStabilities and Thresholds in ANTarctica (INSTANT) program, alongside open-source contributions to the ASPECT mantle convection code. 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 · 2025-07

The ability to precisely measure the connectivity of large networks of neurons, even entire brains, has rapidly grown due to advances in microscopy and image processing. The resulting connectivity diagrams, or "connectomes," promise detailed knowledge that can be used to inform models of brain signaling and, ultimately, the biological basis of intelligent behavior. This project will develop theoretical methods for the analysis of these datasets. As the scale of connectomes grow, such methods will be increasingly important. This project will accelerate the development of approaches that are capable of scaling to large brains and that are robust to the heterogeneity and complexity present in real nervous systems, while also facilitating the recruitment and training of interdisciplinary scientists with strong analytical skills to work with connectome datasets and build new models. There is a relative lack of techniques for exploiting connectomic data for hypothesis generation beyond manual examination of individual connections between previously identified neurons with hypothesized functions. Given the orders of magnitude difference in scale between previously available and more recently released connectome datasets (∼300 neurons in C. elegans vs. ∼140,000 in the adult Drosophila brain, for instance), moving from manual approaches to statistical descriptions and analyses whole-brain connectivity is critical. This project's research involves two approaches: analysis of whole-brain sensorimotor pathways and their relationship to behavior, and structural analysis techniques to identify interpretable low-dimensional organization in connectome data. In the short term, these approaches will accelerate investigations of the connection between structure and function in fruit flies, in particular how diverse behavioral responses across contexts are generated by a common wiring diagram. In the longer term, the principles that we uncover will lay the groundwork for applications to larger datasets in other organisms, including mammals. 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 · 2025-07

This project addresses the critical need to approximate complex statistical models in the era of big data. The investigator aims to develop a comprehensive understanding of when simpler approximations can be effectively used, and to provide rigorous guarantees for their accuracy. The core challenge lies in the trade-off between model complexity (which is often necessary to capture the nuances of large datasets) and computational feasibility. While complex models offer rich descriptions, their computational demands can be prohibitive. The investigator proposes to explore whether simpler, approximated models can retain the essential features of their complex counterparts while significantly reducing computation time. Graduate students will be involved in this research. The research will concretely examine the validity of naive mean-field variational inference in several key areas where complex models are prevalent, which include (i) High-Dimensional Bayesian Regression (linear and logistic), (ii) Latent Dirichlet Allocation (LDA), (iii) Mixed Membership Models, and (iv) Exponential Random Graph Models (ERGMs). For each of these examples, the investigator plans to develop tailored inference methods, and provide rigorous, quantified bounds on the errors introduced by these approximations. This will offer a clear understanding of the trade-off between simplicity and accuracy. The overarching goal is to equip researchers using naive mean-field based variational inference with concrete guidelines on when and under what circumstances such methods can be reliably applied, and when they might fall short. This will advance the principled application of approximate inference techniques in the context of big data analytics, contributing to more efficient and reliable scientific discovery. 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 · 2025-07

Modern scientific data sets—ranging from single-cell RNA sequencing with tens of thousands of genes per patient, to galaxy-survey spectra with millions of stars, to user-item interaction matrices in online platforms—share two features: (i) ultra-high dimensionality and (ii) latent parameters that obey common structural laws (e.g., exchangeability, sparsity, or low-rank dependence). This project tackles both challenges at once. It advances statistical foundations for such problems by (1) providing a new framework to theoretically study empirical Bayes methods in these complex models that learn the latent-parameter distribution directly from the data, and (2) developing cutting-edge unsupervised dimension-reduction techniques that embed the high-dimensional observations into lower-dimensional representations while preserving the essential structure and relationships within the data. Together, these tools will transform ad-hoc prior modeling into an objective, data-driven procedure and yield principled, scalable inference for large-scale applications. Further, collaborations with astronomers will ensure immediate scientific impact, and several of the research directions will shape the Ph.D. dissertation of multiple Columbia graduate students, fostering the next generation of data-science leaders. The project integrates two tightly linked research thrusts. (a) Building on recent advances in nonparametric empirical Bayes, the PI will design flexible empirical Bayes estimators for latent-variable distributions for exchangeable parameters. The central innovation is to merge empirical Bayes ideas with variational approximations in general probabilistic latent-variable models, producing estimators that remain computationally tractable, and achieve optimal risk. (b) Classical linear dimensionality reduction techniques like principal component analysis and multidimensional scaling are often inadequate for datasets that are growing increasingly complex in fields such as genomics, astronomy, and finance. To tackle these challenges, the PI will design non-linear reduction techniques --- combining manifold learning with low-rank factor models ---informed by ideas from optimal transport and algebraic geometry. Together, these advances will deepen statistical theory at the nexus of empirical Bayes, unsupervised learning, and high-dimensional inference. All algorithms will be released as open-source R/Python packages accompanied by tutorials. 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 · 2025-07

The rapid expansion of machine learning (ML) and artificial intelligence (AI) has created an urgent need for tools that enhance trust, and control over these technologies. This project aims to address that need by developing methods that help explain why a model makes certain predictions, improve model tuning, and detect harmful or misleading data—such as intentionally corrupted inputs introduced by adversaries—before they compromise the model’s reliability. A promising strategy for tackling these challenges is to analyze the impact of individual data points or subsets of data by estimating how their removal affects the model’s behavior. This research supports the national interest by advancing scientific understanding of AI, improving the robustness of decision-making systems, and contributing to the development of technologies that align with privacy protections. The project investigates whether it is possible to develop computationally efficient algorithms that approximate the output of a model trained without a given subset of data, without having to retrain the model from scratch. This question is particularly challenging in high-dimensional settings, where the number of features is large relative to the sample size. The project focuses on designing data removal methods that are both scalable and theoretically sound in these regimes. The resulting algorithms will be evaluated in two important application areas: risk estimation and machine unlearning. Through this work, the project aims to lay the foundation for practical tools that improve model interpretability and accountability in complex learning 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 · 2025-07

Many of the most abundant organisms on earth—including humans—are so successful because they cooperate with one another. Although cooperative species tend to be found in harsh environments characterized by drought and unpredictable rainfall, it remains unclear why. Nearly all previous work has established global patterns between social behavior and environment by comparing different species in different habitats. Yet, determining why environmental harshness favors cooperation requires within-species studies across different habitat types rather than just among-species correlations across the globe. By combining field experiments and long-term observations, studies of brain morphology and gene expression, and theoretical modelling, the researchers will establish the largest within-species study of cooperation ever attempted in social wasps that span a range of more than 4000 km across Sub-Saharan Africa. From arid deserts to lush rainforests, Africa’s most widespread wasp Belonogaster juncea is ideally suited to study the links between the environment and cooperation. Using populations in Cameroon, Kenya, and South Africa, the researchers will work at a continental scale to help transform the field of social evolution. Results will be of interest to biologists and social scientists studying cooperation broadly; AI deployment will benefit researchers tracking individuals in complex environments; and our international team (USA, UK, Cameroon, Kenya, South Africa, Taiwan) will promote capacity-building in Africa. The researchers will also provide outreach across continents, enthusing audiences about animal societies and the potential for ‘big picture’ experiments to solve fundamental problems in biology. Paper wasps have long been considered classic models for studying cooperation. The researchers will use the wide-ranging African wasp Belonogaster juncea to test, for the first time in an insect, ideas relating environmental variability to social evolution. By combining (1) field-based behavioral experiments and AI approaches to social network analysis, (2) mechanistic studies of brain transcriptomics, micro-CT-scanning, whole-genome resequencing, and AI approaches to morphological evolution, and (3) evolutionary simulations and game theory models, the researchers will pursue four primary objectives. First, they will test whether aridity amplifies the value of cooperation through field experiments in different environments. Second, they will test whether harsh environments promote peaceful cooperation by making conflict too risky, deploying AI to dissect social networks. Third, they will test the key prediction that seasonal environments drive members of cooperative groups to adopt specialist roles, combining integrative approaches to dissect molecular and morphological phenotypes. Finally, they will examine which environmental conditions likely fostered the origin of cooperation in the ancient ancestors of wasps by developing evolutionary simulations and mathematical models based on real-world data. 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 · 2025-06

Stratocumulus clouds are ubiquitous over large areas of Earth's oceans. Despite their relatively consistent structure compared to other cloud structures, like thunderstorm complexes, there is significant small-scale spatial variability in stratocumulus clouds that impacts further cloud and precipitation development. This project will use an emerging modeling technique and existing cloud observations to identify the sources of spatial variability at the cloud microphysics level. The research team will then identify the best methods to simulate these sources of variability. The broader societal impact of the project would be to improve modeling of clouds which affect Earth’s radiation balance. There is also a substantial educational aspect to the project, enhancing the training of the next generation of atmospheric scientists. The primary objectives of this project are to understand the microphysical sources of spatial variability in low-level stratocumulus clouds, how this spatial variability influences the temporal evolution of the mean cloud and precipitation processes, and how the design of microphysics parameterizations influences the ability to simulate the observed spatial structure of these cloud and precipitation fields. The research team will run Large Eddy Simulations (LES) of drizzling stratocumulus clouds using Cloud Model 1 (CM1) based on observed cases with Lagrangian microphysics. Lagrangian particle-based methods employ particles that are representative of a multiplicity of millions to billions of identical hydrometeors, which get around problems introduced in traditional bulk and bin microphysics schemes. The project will focus on four primary potential sources of microphysical variability resulting in rain: condensation, stochastic collisions, turbulence-enhanced collision kernels, and giant aerosol particles. The results of the simulations and comparison with observations will then be used to train the Bayesian Observationally-constrained Statistical-physical Scheme (BOSS), with the introduction of a stochastic component. Simulations will be run in a purely non-stochastic mode and a fully stochastic mode to address the hypothesis that the non-stochastic BOSS will not be able to simulate the spatial variability whereas the stochastic BOSS will. Further simulations will address the importance of the identified microphysical variability as well as whether BOSS can be run in a single-category mode with increased predictability. 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 · 2025-06

This doctoral dissertation project investigates and collates the interactions of scientific disciplines of nutrition, demography, and epidemiology. It examines how scientific experts deploy new technologies and standards to measure population and disease to control famines, as well as lay knowledge about the effects of the famine on the body. This project offers insights into how scientific knowledge of disease and food changes and is created during a famine. This project applies mixed methods to epidemiological data sets, famine camp records, medical journals, administrative reports, scientific surveys and conference proceedings. The role of diverse actors and organizations, ranging from decision makers, nutrition scientists, surveyors, medical professionals, missionaries, and philanthropies and civic voluntary societies will be examined. Through this analysis, this project traces the ways scientific knowledge conceptualizes famines and its impact on communities and regions. 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 · 2025-06

The oceans play a crucial role in transporting heat and moisture from the equator to the poles. Observations show that the input of freshwater in polar regions from melting glaciers and sea ice has already contributed to weakening ocean circulation. However, our understanding of these processes is limited by the length of observational records; examining previous periods of change can aid this understanding. This project focuses on the Younger Dryas (11.7-12.9 ka), a cold period that abruptly interrupted the Earth’s warming climate during deglaciation. The cause of this cooling is debated, but a leading hypothesis is that freshwater input from glacial lakes to the North Atlantic Ocean slowed ocean circulation. Alternatively, computer models indicate that sea ice export from the Arctic Ocean, traveling along the east coast of Greenland, could have been responsible for triggering the Younger Dryas. Yet, there are currently limited palaeoceanographic data from this region to evaluate this alternate hypothesis. To address this data gap, this project will utilize existing marine sediment cores to reconstruct sea ice and freshwater changes during the Younger Dryas. This research will provide quantitative data to understand what triggered an abrupt slowdown of ocean circulation that resulted in prominent cooling of the Northern Hemisphere. The results will lead to a better understanding of the sensitivity of ocean circulation to the input of freshwater from sea ice export and melt, important for evaluating the potential impact of future melt water release from the Arctic. Additionally, this project was conceived and will be led by a postdoctoral researcher, with support from two established principal investigators. Thus, an early career researcher will develop leadership and supervisory skills, essential to professional development. This project addresses the hypothesis that ‘The Younger Dryas was initiated by the export of Arctic sea ice and freshwater carried by the East Greenland Current’. We will leverage existing marine sediment cores collected from the eastern coast of Greenland. Sea ice and salinity dynamics will be inferred from stable isotope measurements on planktic foraminifera and algal lipids, as well as from sea-ice biomarkers, lipid-based sea-surface temperature reconstructions, sedimentological analyses, and radiocarbon dating. Results will be synthesized with existing data to the north (Fram Strait and the Arctic Ocean) and south (Denmark Strait and the northern North Atlantic), in the context of existing model results, to examine the role sea ice and freshwater played in triggering the Younger Dryas. 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 · 2025-06

A fundamental challenge in environmental science is applying the knowledge that scientists discover at a particular location or time to understanding phenomena occurring at other times and/or locations. In the traditional approach, environmental scientists collect field data and perform experiments at, for example, a particular river basin, and then repeat this process at a different time and/or location to see whether their conclusions generalize. This approach is rigorous, but limited because it is time-, labor-, and cost-intensive; thus there exists relatively sparse ground-collected data across the planet. Another challenge is that intensifying human activities amplify the rates of change in conditions per location and time, so knowledge discovered in the past will likely fail to predict outcomes in the future. The challenge of predicting and ameliorating the effects of environmental change disproportionately affects under-resourced communities, including those most vulnerable to environmental changes that lead to food insecurity and hence greater socioeconomic instability. Traditional Artificial Intelligence (AI) approaches cannot resolve this challenge because they require extensive human input, for example due to the need for labeling ground-collected data or other data layers, such as high-resolution satellite imagery. In this project, on-the-ground human observations and labels are replaced with AI-based discovery from abundantly available, mostly unlabeled visual data, such as that collected from a combination of satellites and other devices. This research proposes a paradigm shift that enables low-cost scaling across many types of images in order to lower the barrier to access of this scientific process. In the process, a novel AI framework is introduced that combines multiple data sources to automatically discover interpretable scientific hypotheses about the cause of ecosystem changes. Together, these approaches will accelerate the ability to identify solutions for the increasing environmental issues faced across our planet. The goal of this project is to develop and validate an AI framework that can use a broad array of image data collected using different sensing modalities (e.g., low-resolution satellite, drone, and internet-posted images) to automate and accelerate the generation of interpretable environmental scientific hypotheses at a planetary scale. An example might be correlating the spatiotemporal prevalence of certain invasive or disease-causing species with presumed causal factors present in the environment. The proposed framework integrates new techniques into foundational models for satellite imagery that can choose intelligently among sparsely-labeled data from different sensor modalities, optimizing between cost and accuracy trade-offs. By coupling this model with self-improving large language models that can both receive and provide interpretable feedback and hypotheses to researchers, this approach goes beyond black-box feature learning, the current state-of-the-art in computer vision. This proposed model will be applied and validated on the task of detecting submerged aquatic vegetation. This task poses a number of technical challenges (e.g., waves, turbidity, weak spectral signal through water) that are more difficult than detecting objects on land surface. Success in this pilot project will demonstrate that this type of model can be easily applied to the terrestrial environment and to tackle even greater grand challenges in environmental science. This award is co-funded by the Directorate for Computer and Information Science and Engineering and by the Directorate for Biological Sciences. 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 · 2025-06

This I-Corps project focuses on the development of automated laboratories for the remote execution of scientific experiments. Automated laboratories allow scientists to submit and manage their experiments over the internet, eliminating the need for a physical presence in a traditional laboratory setting. Currently, traditional laboratories involve significant upfront investments in equipment, resources, and dedicated laboratory spaces, making research costly and limiting accessibility, especially for smaller research teams and educational institutions. By removing these barriers, automated laboratories significantly reduce the cost and complexity of conducting scientific research, enabling participation from researchers in academia, startups, and industry. The widespread adoption of automated laboratories would enhance the reproducibility of scientific research, accelerate innovation across the life sciences, and support national interests by facilitating advancements in healthcare, biotechnology, and pharmaceuticals. This I-Corps project utilizes experiential learning coupled with a first-hand investigation of the industry ecosystem to assess the translation potential of the technology. This solution is based on the development of a cloud-based laboratory platform integrating robotics, automation software, and artificial intelligence-driven systems. The core of this technology involves a programming language specifically designed to describe detailed experimental protocols, which the platform translates into precise instructions executed by robotic instruments and laboratory personnel. Users benefit from unprecedented accuracy, reproducibility, and data transparency, greatly facilitating analysis and collaboration. Furthermore, the integration of advanced artificial intelligence models aids researchers in experimental design, result interpretation, and logistical management. Adoption of this technology will enable researchers to scale their experiments rapidly, conduct parallel experiments efficiently, and significantly streamline the path from hypothesis to discovery. 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 · 2025-06

Nontechnical description: This research project will develop a new way to study unusual behaviors in advanced materials, called moiré materials, made from stacking ultra-thin layers. These moiré materials can create unique "quantum phases," which may lead to future breakthroughs like energy-efficient electronics or powerful quantum computers. However, scientists have not had the right tools to closely examine these tiny layered devices. The research team plans to solve this problem by developing a new kind of light-based tool using terahertz radiation —in between microwave and infrared light. This method, based on a recent breakthrough by the team in generating terahertz radiation, can be used right on the chip and can detect both the energy and structure of these quantum phases. Besides advancing science, the project also includes outreach programs to inspire students of a broad range of backgrounds in New York City to explore careers in science and technology through hands-on research and educational activities. Technical description: The formation of two-dimensional moiré interfaces has created unprecedented opportunities in the exploration of quantum phases of matter, but there is a lack of spectroscopic tools in probing moiré quantum matter, due to a mismatch in sample sizes of moiré devices (~1-10 µm) and the wavelengths of electromagnetic radiation (0.1-3 mm) relevant to the low-energy excitations in these quantum phases. The proposed research aims to fill the critical gap in current research. The research team will develop in-situ terahertz spectroscopy to directly access low energy excitations in moiré quantum matter. This approach is based on the discovery by the team of intense and broadband terahertz generation from the van der Waals ferroelectric semiconductor niobium oxydiiodide with record-setting efficiency. This terahertz emitter can be integrated into van der Waals heterostructures for on-chip near-field terahertz spectroscopy of a target moiré material/device. In addition to the low energy scales characterized by the correlation gaps, this experiment will also quantify the degree of correlation and topologies. The former will be deduced from terahertz photo-conductivity, while the latter will be determined from the connection of band topologies to terahertz optical selection rules or Kerr/Faraday rotation. The research team will i) probe the correlation gaps; ii) measure terahertz photo-conductivity of correlated charge carriers; iii) determine topologies from selection rules of resonant excitation with circularly polarized terahertz field, and iv) determine the dynamics of electronic excitations with optical pump and terahertz probe. Such spectroscopic insights will complement prevalent transport measurements and motivate/guide theoretical efforts in searching and understanding of quantum phases in the broadly tunable moiré material 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.