Purdue University

NIH Research Projects · FY 2025 · 2025-08

Enter the text here that is the new abstract information for your application. This section must be no longer than 30 lines of text. Cryo-electron microscopy (cryo-EM) has become a widely used technique in structural biology for determining 3D structures of biological macromolecules. Despite an increasing number of structures being deposited in public databases like PDB and EMDB, many maps are still determined at 3 Å or worse resolution, posing challenges for structure modeling. Thus, there is a strong need for computational tools to assist in structure modeling and validation, given the increasing use of cryo-EM and cryo-electron tomography (cryo-ET) by scientists. Computational modeling is an integral and indispensable component in structural biology. Similar to the situation with microscope, superior modeling methods have the capacity to extract more accurate structural information from otherwise less informative data from cryo-EM and cryo-ET and provide new insights and opens up new research strategies. The goal of this project is to develop and apply computational methods for biomolecular structural modeling for cryo-electron microscopy. In this project I will substantially expand and enhance the capabilities of structure modeling methods to meet new demands and to improve accuracy and efficiency. This will be achieved by developing a deep learning-based approach that can consider key aspects in structure modeling altogether, including structure heterogeneity, atom detection in the density, structure prediction, interaction between proteins and nucleic acids. For lower-resolution maps, we will introduce a novel approach designed to identify and enhance key structural features within the map, which will significantly improve the accuracy of model building. This approach will also be applied to structure fitting and identification for cryo-ET. Additionally, a method for detecting and modeling small-molecule ligand structures in medium-high resolution EM maps will be developed, which is needed for drug discovery. In the process of building model structures, model validation is of crucial importance. To pinpoint modeling errors in PDB, we have devised a pioneering deep-learning-based quality assessment score, known as DAQ. We intend to expand DAQ's capabilities to identify atom-level errors in protein and nucleic acid structures, while also offering suggested corrections for flawed structure models within the DAQ-Score Database, which presently delivers model validation reports.

NSF Awards · FY 2025 · 2025-08

Biomolecules, proteins and nucleic acids, carry out almost all tasks inside living cells. To understand how such molecules work and how to design new drugs that bind to them, their three-dimensional (3D) structures are crucial. Recently, cryogenic electron microscopy (cryo-EM) has allowed scientists to visualize these biomolecules as 3D volume data, but its resolution is not high enough to capture all the details. Converting those cryo-EM data into precise 3D atomic models remains slow, expensive, and challenging, especially for large complexes. This project aims to develop new artificial-intelligence (AI) tools that will automatically interpret cryo-EM data into accurate 3D molecular structures and check those structures for mistakes. By releasing the software as freely available web computing services and expanding an open database of quality assessments, many academic laboratories, biotech companies, and pharmaceutical companies can enhance their research and development. By this project, hands-on workshops on developed tools and outreach activities will be conducted. Faster, more reliable cryo-EM modeling will accelerate drug discovery, enhance numerous structural biology research efforts, and lead to new applications of AI in the biological field. This project will pursue three closely related objectives: (1) it will develop a deep learning-based method capable of constructing protein-DNA/RNA complex structures. The proposed architecture will integrate advanced deep learning architectures, making the process more accurate and scalable for large protein-DNA/RNA complexes. Additionally, a novel method will be developed to identify unknown proteins and nucleic acids within cryo-EM maps with high accuracy. (2) The model quality assessment score framework will be expanded in two directions: The score will assess backbone and side-chain errors in high resolution EM maps and also scores that evaluate molecules other than proteins will be developed. (3) For challenging medium to low-resolution cryo-EM maps, a new biomolecular modeling and assembly method will be developed. It will sharpen density with a diffusion model, convert it to a backbone point cloud, and fit the local structure of AlphaFold models using advanced point-cloud registration techniques, followed by clustering and real-space refinement. Expected outcomes include open-source software, a publicly accessible online computing web service, and an expanded quality-assessment database. These outcomes will overcome the current limitations in biomolecular modeling for cryo-EM 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 · 2025-08

Powered lower limb prostheses have made significant technological progress, yet they continue to face major challenges in navigating real-world environments. Individuals with lower limb amputation often struggle with activities involving transitions between surfaces such as grass, gravel, or slopes, where current prosthetic controllers are limited in providing the necessary stability and agility. This Smart and Connected Health (SCH) project seeks to address that challenge by advancing the control strategies used in robotic ankle-foot prostheses. By understanding how humans adapt their movement when walking on different terrains, the research will inform the development of next-generation prostheses that can proactively and reactively adjust to changing environments. The goal is to enhance the independence, safety, and quality of life of individuals who rely on powered prostheses. In doing so, this work contributes to the national interest by promoting the progress of science and engineering, improving public health and welfare, and inspiring future innovation in assistive technology. The project seeks top support education by engaging students in hands-on research experiences that foster interest in robotics, biomechanics, and assistive technologies. The project will investigate the sensorimotor mechanisms that enable humans to walk dynamically over uneven and unpredictable terrain. These insights will be used to develop and validate a novel, multi-level control architecture for powered prosthetic limbs. The approach integrates proactive strategies based on predictive terrain recognition with reactive feedback-based stabilization. The control framework will be tested in real-world scenarios with individuals with limb loss to assess performance and identify failure modes. The research seeks to improve understanding of how to merge human neuromechanical signals with environmental feedback in a unified, robust control system. Outcomes look to include new methods for real-time adaptation in wearable robotics, advancements in the field of human-in-the-loop control, and broader applications in rehabilitation robotics and mobility assistive devices. Through rigorous experimental evaluation and community engagement, the project seeks to redefine the future of lower limb prosthetic function. 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

Drinking water is not sterile. Many microbes thrive in biofilms along water pipe walls. While most of these microbes do not cause harm, some are pathogens that can cause respiratory infections. These pathogens cost more than $2.4 billion in healthcare costs each year. Water distributors and building owners try to kill biofilm bacteria with disinfectants and high temperatures, but biofilms are resilient. The goal of this career project is to engineer a biofilm that promotes good bacteria that can outcompete unwanted ones. If successful, this strategy to work with biofilms, instead of against them, will lead to more stable drinking water quality. It will also be more sustainable, with lower energy and material costs. Understanding more about biofilms in our water pipes will improve guidance that makes drinking water safe. Educational and outreach activities will strengthen the Nation’s STEM workforce and empower the public to take control of their own drinking water quality. Most microbes in drinking water systems reside in biofilms along pipe walls. While these microbes are largely benign, drinking water pathogens associated with the immunocompromised, or opportunistic pathogens, thrive in biofilms. Legionella pneumophila, Pseudomonas aeruginosa, and non-tuberculous Mycobacteria (NTM) are estimated to cost more than $2.4 billion in healthcare costs each year. Current strategies to control these pathogens in plumbing systems, if effective, are unsustainable, with high energy, maintenance, and material costs. Even with disinfectant, biofilm formation in plumbing is largely inevitable. New strategies are required to control these biofilms and reduce exposure to pathogens from drinking water. An approach embracing biology can be more sustainable and provide greater stability in plumbing systems. This project will engineer drinking water biofilms to have long-term biological stability and resistance to invasion, especially by pathogens. The scalable approach will establish a competitive model microbial consortium specific to drinking water. At the bench-scale, coupon experiments and a cell printing method will be used to control initial biofilm colonization. A unique Plumbing Testing Facility will be used to assess long-term biofilm stability at real scale. Finally, the same facility and pure culture experiments will be used to explore and validate models of biofilm-water interactions. This approach considers the unique reality of plumbing, with high surface area to volume ratios and intermittent flow. Using a community ecology framework allows connection of this applied project to broader understanding of dispersal and selection processes. Educational activities will help educate the next generation of students in public health, biological engineering, and environmental engineering. They will also educate the public on drinking water and biofilm engineering to increase scientific literacy. 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

Reinforcement learning (RL) is a promising approach for enabling machines, such as robots or cars, to make decisions in complex and unpredictable environments. Examples of these are robots that can run or autonomous cars that can navigate cluttered streets. To make these algorithms work, people use simulation. The problem is that in practice, these robots struggle to solve similar challenges in the real-world, due to the lack of controllability in these applications. The more realistic the environment, the more data that are required, and the more time that is needed for such robots to work effectively. These challenges are made even more difficult by what is known as the sim(ulatio)n-to-real gap, which refers to the problem of applying what has been learned in computer simulations to real-life situations where conditions are different and less predictable. These limitations restrict the use of the algorithms and robots in important areas where collecting data is expensive, risky, or impractical, such as healthcare robotics or emergency response systems. This project will address these key limitations by eliminating reliance on simulators altogether and developing reinforcement learning methods that can be trained directly in real-world settings. The project will create new theoretical frameworks and algorithms that integrate foundation models, which are large, pre-trained artificial intelligence models, into reinforcement learning. These models provide built-in knowledge about the world, enabling learning systems to acquire new skills more quickly, perform better in unfamiliar situations, and be deployed more rapidly in real-world environments. The research will focus on three core areas. First, it will establish formal methods for incorporating knowledge from foundation models into the state and action spaces of reinforcement learning agents, allowing them to leverage high-level abstractions and prior knowledge to inform decision-making. Second, it will derive theoretical analyses and provide performance guarantees, including bounds on sample complexity, generalization capability, and convergence rates, to demonstrate how these enhanced agents can learn more efficiently with fewer interactions in complex, real-world environments. Third, it will design new reinforcement learning algorithms that integrate foundation models to improve reward modeling, exploration strategies, and policy optimization, ensuring stability and robustness during both training and deployment. 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 study focuses on how engagement scholarship can impact research design, practices and approaches. The project traces the trajectories of robotics research as it is translated from the research lab to medical and commercial markets. The findings of this study inform robotics research practices as well as contribute to increased outcomes for local, state, and national audiences and markets of robotic assisted technologies. The project designs and implements a series of professional development activities focused on needs assessment and translation planning with targeted audiences for robotics researchers. The research team tracks disciplinary effects of this intervention on the publications and research activities of the participating robotics researchers. Participating researchers are assessed for knowledge development through pre- and post- interviews. The findings of the study contribute to understanding how researchers who develop robotic assisted technologies can improve their research outcomes through engaged practices. This project was funded by the Science and Technology Studies Program and the Ethical and Responsible Research Program in the Social, Behavioral and Economic Sciences Directorate. 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 addresses a central challenge in artificial intelligence (AI) and data science: how to make accurate and reliable decisions when dealing with complex, discrete data; that is, data made up of distinct elements like words, networks, or molecular structures. Many important tasks in science, engineering, and everyday technology rely on the ability to reason under uncertainty in these settings. However, current algorithms often struggle to efficiently explore the vast space of possibilities such data can represent. This research will develop a new class of sampling algorithms designed to be faster, more scalable, and more statistically reliable, making it easier for machine learning systems to handle discrete data effectively. These advances will support the development of trustworthy AI systems, improve scientific simulations, and enable more controllable generative models in areas like drug design, recommendation systems, and natural language processing. By releasing open-source tools and training a diverse group of students and researchers, the project also contributes to the broader scientific community and helps build a skilled workforce. In doing so, it supports the national interest by advancing the progress of science, promoting innovation, and enhancing economic and societal well-being. This project develops a new framework for discrete Markov chain Monte Carlo (MCMC) algorithms that leverages gradients. The proposed gradient-based discrete MCMC (GD-MCMC) approach provides more informed exploration and significantly improves convergence compared to traditional methods. The research plan is structured into three thrusts. The first thrust focuses on developing algorithmic innovations for MCMC in non-log-concave and non-smooth discrete distributions. The second thrust establishes theoretical foundations by providing convergence guarantees for both log-concave and non-log-concave settings. The third thrust demonstrates practical impact in both machine learning and scientific domains, with applications in discrete generative models and molecular optimization. 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 rapid advancements in neuroscience, robotics, and computer systems have underscored the vital interactions between mechanical and computational systems in shaping behavior. In natural systems, such as those found in animals, the brain and body must collaborate effectively for the successful navigation of a complex environment. The brain contributes computational intelligence, while the body provides mechanical intelligence. Integrating these elements—computational and mechanical intelligence—into the concept of mechano-computation represents a frontier in both robotics and neuroscience research. Progress in this field necessitates interdisciplinary communication and collaboration across various scientific domains. To propel this promising field forward, the Mechano-computation for Expanding Scientific Horizons (MESH) Network aims to unite diverse researchers from robotics, mechanics, materials science, neuroscience, information theory, biology, engineering design, and applied mathematics. Through workshops, travel grants, and the facilitation of collaborative projects, this network seeks to stimulate interdisciplinary dialogue, develop rigorous metrics for assessing autonomous systems, train the next generation of researchers, and push the boundaries of research in all areas of mechano-computation. By establishing a centralized resource for sharing findings, benchmarks, and methodologies, this network of researchers can accelerate innovation and position the United States as a leader in this transformative field, laying the groundwork for enhanced robotic systems in healthcare, agriculture, forestry, national security, and beyond. It may be argued that the full potential of robotics will not be realized until an intelligent physical body is purposefully designed from the outset, with careful consideration of both the available computational intelligence and the affordances the body can provide—affordances that, if appropriately leveraged, can offload and simplify computational demands by enabling efficient, embodied solutions to complex tasks. The Mechano-computation for Expanding Scientific Horizons (MESH) Network will bring together leading experts to tackle these critical challenges in autonomous systems through the integration of mechanical and computational intelligence. Creating intentional mechano-computation will enhance the design and control of autonomous systems, making them more efficient and explainable, and it will contribute to the development of innovative materials, mechanisms, and control strategies, pushing the boundaries of current research. We anticipate five key outcomes as a result of the formation of the MESH Network: (1) A comprehensive theoretical framework and standardized metrics for mechano-computation; (2) Improved interdisciplinary collaboration and communication among researchers; (3) Long-term interactions among network members and early-career researchers, including nurturing graduate students trained at the intersection of disciplines; (4) Sharing of innovative materials, mechanisms, and control strategies; (5) Practical demonstrations by network participants of mechano-computation systems addressing societal and environmental challenges. The network will accomplish these outcomes through tasks that build online repositories of network critical technical and organizational information, in-person events to broaden discussion and collaboration, online communities, and targeted support for bringing in new collaborative research areas. This project is supported by the Dynamics, Control, and System Diagnostics (DCSD), the Engineering Design and Systems Engineering (EDSE) and the Mechanics of Materials (MoMs) programs of the Division of Civil, Mechanical, and Manufacturing Innovation (CMMI) in 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 · 2025-08

With the support of the Chemical Catalysis program in the Division of Chemistry, Professor Hibbitts of Purdue University and Professor Plaisance of Louisiana State University are developing new computational models to study how chemical reactions on catalyst surfaces are influenced by the presence of other species on the surface. Reactions on solid catalysts occur on surfaces that range from being nearly empty to very crowded, depending on the reaction conditions such as temperature and pressure. Adsorbed species on crowded surfaces form a layer that can influence reaction rates, thus impacting the reactivity, selectivity, and stability of a catalyst. Currently, theoretical methods to interrogate the effects of these layers are cumbersome and time consuming, and the federal funds provided through this project will establish new models that will greatly facilitate including these layer effects. This model will describe the layer using an “implicit” model whereby it is treated in a continuum approach rather than through a more complex and costly “explicit” model that treats the layer using more expensive quantum chemical methods. This project will further our fundamental understanding and ability to control catalytic reactions, which is key to chemical transformations in all industries. In addition to these research impacts, educational research opportunities will be extended to high school and undergraduate students through merit-based outreach programs at both Purdue and LSU that train students in theoretical chemistry methods. With the support of the Chemical Catalysis program in the Division of Chemistry, Professor Hibbitts of Purdue University and Professor Plaisance of Louisiana State University are developing an implicit model to study the effects of adlayers on heterogeneous catalyst surfaces within density functional theory (DFT) calculations. Catalyzed reactions often occur on crowded surfaces, particularly at high pressures or in condensed phases. These adlayers behave similarly to a two-dimensional liquid in which the interactions with co-adsorbed intermediates and transition states are analogous to interactions between solutes and the surrounding solvent in liquid-phase reactions. Quantum chemical methods like DFT have long modeled solvent effects using two strategies: explicit models where the solvent molecules are treated at the same quantum chemical level as the solute, and implicit models in which the solvent is represented as a continuum field. Here, we will develop a novel implicit adlayer model that will account for (1) the influence of the adlayer on the binding properties of the catalyst surface, (2) the interaction of intermediates and transition states with the surrounding adlayer, and (3) thermodynamic effects associated with the displacement of the adlayer from the catalyst surface by such intermediates and transition states. These models will be parameterized by DFT calculations of co-adsorbate effects on both single-crystal and nanoparticle catalyst surfaces. Four catalytic reactions, previously studied using explicit adlayer models, will be re-examined using the implicit adlayer models to check for their accuracy: alkane hydrogenolysis on H-covered surfaces, CO hydrogenation on CO-covered surfaces, methane activation on O-covered surfaces, and hydrodechlorination on Cl-covered surfaces. These reactions are relevant to many traditional and emergent catalyst applications and are chosen as each adlayer offers unique challenges to be overcome. This project will further our understanding of how adlayers influence catalytic reactions and develop an implicit adlayer model that will enable researchers to rapidly establish these effects using a tunable approach. 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 is focused on inverse problems, where one seeks to recover the parameters of unknown media from remote measurements. First, the investigator will study the recovery of a Lorentzian metric, up to a gauge group of transformations, from measurements of the way light or positive mass particles propagate as observed on a timelike boundary. In cosmology, this means recovery of spacetime from remote observations. It has applications to probing moving media with acoustic waves as well. The measurements could be either arrival times and directions of rays or more generally, the arriving wave itself. The second problem is to recover the underlying Riemannian geometry in a bounded domain, say in three dimensions, from the area of the minimal surfaces attached at various loops on the boundary. This problem arises in relativity, precisely in the anti-de Sitter/conformal field theory correspondence, sometimes called the holographic duality in physics. The stability of the recovery will be studied as well, i.e., not so sensitive to small errors in the data. The third type of problem to be studied is recovery of the nonlinear parameters of media from the way light or sound, etc., propagate. In particular, the investigator will show that one can achieve a two-wave interaction in nonlinear wave propagation, which does not fit within the conventional framework. Graduate students and postdoctoral researchers will be mentored and trained as part of the project. The project is primarily in the area of inverse problems in Lorentzian geometry, the inverse problem for minimal surfaces, and in nonlinear wave propagation and related inverse problems. More concretely, the investigator studies the recovery of a Lorentzian metric, up to a gauge group of transformations, from measurements of the way light or positive mass particles propagate as observed on a timelike boundary. This is called lens/scattering rigidity. The linearization is a tensorial X-ray transform restricted to lightlike or timelike geodesics. What makes the nonlinear and the linear problems fundamentally different from their Riemannian version is that the linear one loses the ellipticity of the Riemannian case. In particular, stability is lost. A version of this problem when the whole wave is observed, is studied as well using the hyperbolic Dirichlet-to-Neumann map as data. The minimal surfaces inverse problem on a compact Riemannian manifold with boundary asks whether one can recover a Riemannian metric from the knowledge of the areas of the minimal surfaces with prescribed boundary intersections (say, 1D loops in 3D). The investigator plans to investigate stability as well. Finally, it will be shown that one can force a two-wave interaction in nonlinear wave propagation. This interaction is used to recover the nonlinear parameters and the geometry, locally, involved in nonlinear hyperbolic partial differential equations. 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 complex tasks are hard because they make use of multiple kinds (modes) of data at once, such as replying to a question based on both the question-asker’s words and gestures, or generating a video based on a screenplay, a directing style, and knowledge about the audience. Many artificial intelligence approaches have been attempted on these multimodal tasks. While conventional deep neural networks struggle with them, multi-stream architectures are an emerging approach that has been shown to perform better. However, despite their potential, the broader adoption and advancement of multi-stream methods and models are limited by gaps in existing algorithms and their implementations, as well as a lack of foundational design knowledge. This project aims to address these shortcomings by integrating modern developments in deep learning with efficient algorithmic implementations to produce a library of state-of-the-art multi-stream models. The anticipated outcomes include: (1) a suite of high-performance multi-stream foundation models with applications to object detection, text-based image segmentation, and audio-video analysis; (2) optimized algorithms that enhance the efficiency of their internal mechanisms; (3) and new foundational knowledge to guide the design of future multi-stream systems. These outcomes are expected to advance the scientific frontier of deep learning while simultaneously supporting the community with cyberinfrastructure that includes publicly available models and high-quality software implementations. This project will explore two major directions that will advance cyberinfrastructure for multi-stream architectures: (1) the design, training, and scaling of novel multi-stream architectures, and (2) the development of efficient implementations for their core components based on hardware-software co-design. Combining these approaches, we will synthesize theoretical and algorithmic improvements with efficient implementations, thus providing robust cyberinfrastructure for future work. Evaluations will be conducted across diverse tasks, including classification and generative modeling, and on a range of hardware platforms, from edge devices to desktop-grade systems and high-performance computing systems. Comparisons with existing state-of-the-art neural networks will enable us to quantify improvements in both accuracy and efficiency, with particular attention to expanding the accuracy-efficiency Pareto frontier. All datasets, pretrained models, and optimized implementations, such as custom kernels and operators, will be made publicly available to maximize impact and reproducibility. 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

Cloud and network providers must allocate and manage their compute and network resources efficiently and fairly to ensure stringent service level objectives are met. Examples include managing Wide Area Network (WAN) bandwidth to meet traffic needs and managing clusters of Graphics Processing Units (GPUs) for running Artificial Intelligence (AI) training jobs. The management of production networks and systems is a black art today, dominated by domain-specific heuristics that require constant fine-tuning as requirements evolve, and which fall short of performance in ways that are not easy to analyze. This project is developing BONSAI (Beyond-Optimization Network System Allocation Intelligence), a principled framework based on Machine Learning (ML) that can enable high decision quality for a wide range of network resource allocation problems in complex network and cloud environments that are hard to model precisely, and where inaccurate predictions about future traffic patterns or workloads is the norm. The project is developing custom neural architectures inspired by network optimization models for multi-criteria resource allocation problems. The neural architectures will be designed to show resilience to scenarios beyond their training data (e.g., flash crowds, hardware failures) by careful alignment with the optimization models they aim to enhance and adopt a framework that reacts to a wide variety of input transformations with awareness. The project is developing techniques by which neural models can learn and adapt to real-world sources of feedback which are non-differentiable, and novel alignment approaches that will enable real-time validation of the solution suggested by the neural architecture. The project is demonstrating these ideas in the context of important and challenging domains such as traffic engineering in Wide Area Networks and scheduling distributed AI training jobs in shared GPU clusters. The team consists of researchers with complementary expertise in networking, machine learning, and optimization, and results will be disseminated to researchers in these communities. The team will also collaborate with and disseminate results to industry and network operators for real-world validations and applications of their research. The research will benefit the networking and IT industry by taking a major step forward towards principled resource allocation for complex real-world tasks. The project will involve Ph.D., master’s and undergraduate students, and will lead to the creation of a new graduate class on ML-Driven Computer Networking. Project results will be made publicly available at: https://purdue-isl.github.io/projects_pages/Self-correcting-ML Results will be available throughout the project period and for at least three years after. 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 use of Fourier series is ubiquitous throughout much of science and engineering, allowing approximately periodic phenomena to be understood and analyzed efficiently for critical applications. The Hardy-Littlewood (circle) method applies Fourier series to count solutions of Diophantine equations. Studied since ancient times, these are equations to be solved in integers, whose properties remain a central focus of modern research. In general, the circle method uses subtle aspects of Fourier series in the guise of exponential sums, tools that contribute to tests for equidistribution (apparent "randomness") of number theoretic sequences used in computer science and cryptography. Despite the recent renaissance in the Hardy-Littlewood method, the current understanding of its nuances remains plagued by basic mysteries that stand in the way of ambitious applications in analytic number theory that would be transformative in nature. The Principal Investigator will explore these frontiers of the subject, delivering progress on model problems that should enhance our understanding of the most difficult regions of the Fourier space in applications. Informed by these advances, novel approaches, some only recently established and others speculative in nature, will be employed to address fundamental problems concerning Diophantine equations, congruences and equidistribution, opening new avenues for future research. This project will also involve training graduate students in this promising new technology, and work will continue on a text making this research available to the wider mathematical community.   Two great mysteries of the Hardy-Littlewood method concern the nature of contributions in circle method integrals of points with large height, and the relation of special subvarieties of solutions to the associated Fourier analysis. In mean values of exponential sums over polynomials of larger degree, the provenance of points on low degree subvarieties contained in the associated hypersurfaces remains speculative. Three goals will be addressed in this project that address these mysteries. First, height-scaling techniques will be applied in the circle and related methods to obtain new conclusions concerning the representation of arithmetic sequences by higher degree forms. This will provide novel approaches to understanding the representation of integers in such sequences as the squarefree numbers by polynomials of large degree. Second, novel generalizations of the delta-function technique will be investigated, offering access to Diophantine problems of higher degree. This method offers a potential tool for the investigation of the contribution of points of medium circle method height in applications of the Hardy-Littlewood method. Subject to availability of time and resources, a third goal will be to apply congruences over p-adic function fields to derive mean value estimates for exponential sums applicable to conventional Diophantine problems. 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 paint, ink, and pharmaceutical products are suspensions of microscopic particles in a water-based liquid. The performance of these and similar products can be compromised if the particles agglomerate or stick together or if they sediment over long periods of time. Usually, a chemical is added to the mixture to act as a “dispersant,” which prevents or slows down agglomeration and settling of the particles. This project focuses on certain dispersants that form “vesicles,” which are balloon-like flexible spherical containers in which a microscopic membrane completely encloses a water-like region. Dispersants that form vesicles can efficiently prevent particle agglomeration and sedimentation for a wide variety of particle types and sizes. For example, dispersants that form vesicles in various inks has been shown to improve their performance and reduce production costs. The project will use experimental, theoretical, and computational methods to better understand how the formation of vesicles stabilizes suspensions. The project will also evaluate several other dispersants to tailor their effectiveness to specific particle types. Finally, the project will allow undergraduate students to participate in a unique summer research program and will provide interesting and practical examples that can be easily introduced into various engineering courses. This project focuses on the physicochemical mechanisms underlying the use of vesicles and liposomes to control the stability of various suspensions. A combination of experimental, theoretical and computational methods will be used to generate insights into vesicle and liposome dispersion behavior and improve the understanding of these types of complex fluids. The project will screen, evaluate, and select those surfactants which form liposomes and vesicles and establish rigorous property-stabilization relationships. Innovative methods, such as static light scattering and small-angle x-Ray scattering (SAXS), will be employed to determine the intra-vesicle and inter-vesicle microstructures that form and how they prevent particles from settling. This project will also yield a robust and well-tested set of tools for future studies of technologically relevant colloidal suspensions and dispersions. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NIH Research Projects · FY 2025 · 2025-07

Innovative and integrative training in toxicology is essential to produce the next generation of scientists who can translationally address the role of environmental and occupational exposures in adverse human health outcomes and identify interventions to prevent, reduce, or treat toxicity. This proposed training program seeks to provide knowledge that spans biological scale and that goes beyond a trainee’s specific chosen area of research focus. It aims to prepare trainees for effective engagement beyond their areas of specialty to allow success in collaborative translational professional endeavors. The Toxicology Training Program at Purdue University aims to innovatively train PhD graduate students in bidirectional translation. To accomplish this goal, we will frame the Adverse Outcome Pathway (AOP) concept as a training template. AOPs are presented linearly from low (molecular) to high (phenotype) biological scale. Our program aims to use the AOP framework for transformative training across biological scale to connect basic, applied, and clinical advances well beyond current integration and in both directions of biological scale. The Toxicology Program at Purdue University is uniquely qualified to undertake this effort. First, Purdue Toxicology Faculty are world leaders in the field as evidenced by funding and publication records and internationally recognized leadership in toxicology organizations, grant review panels, roles in regulatory decisions and scientific journal boards. Moreover, our toxicology-related faculty (also preceptors) encompass disciplines (i.e., chemical engineering, epidemiology, exposure assessment, imaging) highly amenable to fostering our proposed excellence in training. Second, Purdue toxicology research programs themselves span multiple biological scales, linking primary mechanisms of toxic action in laboratory studies to human adverse outcomes. Third, our technical capabilities are amongst the world’s best suited to accomplish this training. Fourth, our core facilities are also training centers, where trainees learn how to conduct experiments and analyze data, alongside Ph.D. level directors and their staff – trainees will leave the program as technical and theoretical experts in core facility equipment. Fifth, our T32 curriculum is directly designed to address the overarching goal of training in the fundamentals of toxicology and disciplines that foster a coalescence of excellence in bidirectional translation across biological scale. Sixth, the Purdue Toxicology Program already has a long track record of producing independent toxicology leaders in academia, consulting, government, and industry. Our proposed T32 will build upon well-established strengths present in the Purdue Toxicology Program and produce trainees with both breadth and depth in toxicant-biological interactions across biological scale, where every trainee will be uniquely qualified to identify and link primary mechanisms of toxic action to phenotypic changes important in individual and population risk to AOs. Our T32 trainees will be uniquely positioned to address the most important problems facing all sectors of toxicology and rapidly pivot to address emergent problems in the field.

NIH Research Projects · FY 2025 · 2025-07

Project Summary The accurate prediction of B-cell epitopes is essential for disease analysis, diagnostic tests, and vaccine design, yet it lags behind T-cell epitope prediction in precision. This is due to the complex nature of B-cell epitopes and the undersampling of mapped antibody-antigen structural data. To address this, we propose calculating the intensity of immune selection as an indicator of immunodominance because immune response-driven selection leaves detectable genetic signatures around common regions targeted by antibodies. Our preliminary data suggests that surface-adapted immune selection statistics recover common epitope sites in Sars-Cov-2 Spike, Influenza HA1, and malaria antigens. We have set out two primary aims to enhance B-cell epitope prediction. Aim 1: we plan to build a novel and comprehensive antigen database from 68 human pathogens, which maps 3D immune selection profiles onto the surfaces of antigens from common pathogens. Antigens will be selected from IEDB and their underlying population- level variation will be extracted from public genomic resources. Population genetics scores, such as Tajima's D and BetaScan, will be calculated on antigen surfaces. The resulting database will be deployed online for easy public access. Aim 2: we will develop a B-cell epitope predictor with two innovations. First, the training output comes from the normalized selection statistics of antigen surface instead of relying on antigen-antibody structures. Second, structural features for training inputs will be encoded through the Holographic-CNN model. The predictor's efficacy will then be compared against the state-of-the-art models using a distinct test set of experimentally resolved antigen-antibody structures. Upon completion, our predictor is expected to substantially elevate the predictive power of B-cell epitopes. Our antigen selection database will provide unparalleled new information for various research purposes of antigen evolution. These tools will be instrumental for reverse vaccinology, especially the design of epitope-based vaccines and the evaluation of the potential effectiveness of immunological interventions.

NIH Research Projects · FY 2025 · 2025-07

Summary Lipid nanoparticle (LNP)-based formulations are widely used for delivering macromolecular therapeutics, including mRNA. Biologics account for ~55-60% of the total current pharmaceutical product market, with all trends pointing to continued increases in market share. However, transfection efficiency of mRNA is severely hindered by various loss mechanisms, including rapid clearance, suboptimal cellular uptake, and incomplete endosomal escape. These processes reduce the effective delivery of mRNA to only a small fraction (~1-2%) of the administered dose. As a consequence, comparatively subtle patient-to-patient differences in loss can translate to large variability in therapeutic dosing, with corresponding variability in efficacy and side-effects. To improve transfection yields and better understand the intracellular barriers to mRNA delivery, particularly during the key step of endosomal escape, there is a pressing need for new tools capable of providing chemically selective, nanoscale insights in the intracellular fates of LNPs and their cargos within live cells. The goal of this project is to develop fluorescence-detected photothermal infrared (F-PTIR) microscopy to track the intracellular processes governing mRNA release from LNPs in real-time and with ultra-high spatial resolution, well below the optical diffraction limit. In brief, fluorescence from labelled mRNA will serve as a local temperature sensor. Upon IR absorption from the surrounding medium, local transient temperature increases result in corresponding reductions in fluorescence quantum yield. Change in fluorescence intensity as the IR wavelength is tuned enables IR absorption spectroscopy with a spatial resolution set by fluorescence imaging and heat transfer. Embedding fluorescently labeled mRNA within deuterated LNPs will yield IR spectra dominated by the CD stretching modes of isolated LNPs. Endosomal uptake will be tracked both by the physical position of fluorescence within single cells and by the changes in the lipid vibrational spectra (e.g., addition of CH-stretching modes from native lipids in endosome membranes). Tight timing control in combination with heat-transport modeling will be used to quantify distances between the fluorescence reporter and different IR absorbers. The relatively rare subset of internalized LNPs capable of releasing cargo intact into the cytosol will be identified by their corresponding change in local microenvironment (e.g., loss of CD stretches for mRNA solubilized within the cytosol) together with their 3D position and mobility measured by single molecule localization microscopy (SMLM). Once developed and validated, the proposed instrumentation can support informed optimization of therapeutic nanoparticle formulations designed to promote dosing yield and therapeutic efficacy.

NSF Awards · FY 2025 · 2025-07

This project focuses on the development of speech-based artificial intelligence (AI) agents that are carefully designed using proven educational (pedagogical) principles. These AI agents act as assistive tools to help students improve their foreign language skills and enable extended, realistic conversational practice outside the classroom. Furthermore, this innovation can reduce instructors’ preparation time by up to 80% while increasing learners’ speaking practice time by as much as 90%. This technology has the potential to significantly strengthen foreign language acquisition and increase access to foreign language instruction, supporting skills that are vital to the strategic interests of the U.S. 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 technology addresses foreign language learning obstacles related to limited class time, student anxiety, and limited opportunities for practice, which hinder the development of foreign language skills necessary for basic conversation. Unlike existing tools, this technology utilizes recent innovations in generative artificial intelligence (GenAI) including large language models (LLMs) and large speech models (speech-to-text “STT” and text-to-speech “TTS” technologies) to design speech-based, language-agnostic agents that align with and augment classroom pedagogy. This technology’s low cost and ease of integration represent a potentially important advance by providing cost-effective, scalable solutions for U.S. education institutions and students mastering new language skills. 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 creates StreamCI, an open-source streaming data platform designed to help researchers and users across industries to more effectively harness massive sensor data streams using modern data analysis methods such as artificial intelligence (AI) and machine learning (ML). StreamCI simplifies the collection, management, and analysis of sensor data streams in a user-friendly, cloud-based system accessible to the broad research community. By lowering technical barriers and making data AI-ready, StreamCI empowers domain experts to build intelligent and responsive applications that drive faster discoveries and more effective solutions. The platform also serves as an open educational resource, training the next generation of data- and AI-savvy researchers to create data-driven solutions for critical societal needs. StreamCI is designed to streamline the entire workflow for sensor data streams, from capturing raw sensor data to processing it at various levels of fidelity, anonymization, and transformation; and to apply suitable ML methods across different modalities and enable Findable, Accessible, Interoperable, and Reusable (FAIR) data sharing for research reproducibility and cross-domain science. The core innovations of the project include high-level data abstractions that allow researchers to combine data streams across sources; AI-readiness through novel tools for data preparation and canonical and customized data pipelines; low-code application development and integration of powerful tools; and scalable, seamless service deployment and delivery. StreamCI is used across a wide range of domains, including energy, manufacturing, transportation, agriculture, ecology, audiology, and education, demonstrating its broad applicability and versatility. StreamCI is deployed as an open-source, cloud-based software-as-a-service. Its open-source release allows researchers to set up and operate StreamCI on their local computers. 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 aims to serve the national interest by modernizing the teaching of physics at the college level. Specifically, the project will enable college instructors to integrate specialized computational physics methods into their classes, and to rate their students' learning against a standard scale. Teaching computational physics is important because these methods are quickly becoming essential for physicists and other scientists around the world. This work will help colleges and universities keep up with these changes. Rating students against a standard scale will enable individual instructors and departments to determine how effective their teaching efforts are, and to make improvements that further benefit their students. Adding computational methods to college classes will enhance the effectiveness of the U.S. scientific workforce, one of the key goals of NSF and the IUSE program. This Engaged Student Learning: Level II project is the first effort to establish standards (and training materials for using the standards) designed specifically for evaluating students' achievement of seven essential learning goals in computational physics. The goals of this project are to expand and improve the teaching of computational physics at five universities in the midwestern U.S: Indiana University Indianapolis, Bradley University, Purdue University, University of Indianapolis, and University of Wisconsin - Stout. The project team will develop seven student learning objectives: 1) use generative AI effectively and ethically, 2) read, understand, and modify existing code, 3) apply common computational tools, 4) test code, 5) explore physics, 6) write clear code, and 7) communicate physics. Each partner institution will have a specific role and responsibility with respect to developing these learning objectives. The project will also develop, test, and improve rubrics related to these objectives that instructors can use to rate students' learning of these methods. Instructors are accustomed to evaluating students within a single class when they assign grades, but this type of rating does not give information about how a student has progressed over years. To measure progress, it is necessary to rate students against objective standards, therefore this project will also produce documents and procedures a department can use to help its members learn to use the rubrics to rate students objectively. This project will also study how students' development as rated by instructors compares to their own view of how much they have learned. Project work and findings will be disseminated through publications, presentations, and workshops for faculty. The NSF IUSE: EDU Program supports research and development projects to improve the effectiveness of STEM education for all students. Through the Engaged Student Learning track, the program supports the creation, exploration, and implementation of promising practices and tools. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Awards · FY 2025 · 2025-07

The transportation sector is one of the largest producers of greenhouse gas emissions (GHG). Electric vehicles (EVs) are a key technology to reduce pollution from transportation and demand for EVs has been increasing in response to consumer health and environmental concerns, greater vehicle choice and safety features, improved battery capacity, and cost savings. The adoption of EVs and deployment of charging facilities both require advanced system-level integration and awareness, including models of the interactions and operations of the transportation and electric power systems. This IRES project supports U.S. graduate and undergraduate students to engage in a 10-week summer international research experience related to transportation electrification at the University of Aukland (UoA) in New Zealand, which builds upon an existing partnership between the Advancing Sustainability in Powered Infrastructure and Roadway Electrification (ASPIRE) Engineering Research Center (ERC) and UoA’s Transportation Research Center, to accelerate the adoption of EVs through wireless technologies. The Cultural Awareness and Research Experiences for Electric Vehicles (CARE-EV) IRES project prepares students to engage in research on a global scale that utilizes core international partner the University of Auckland (UoA) in New Zealand. This objective is accomplished through an international partnership between ASPIRE and the UoA, where the theme of the program and collaboration is “Advancing charging technology and infrastructure for a sustainable transportation future.” The CARE-EV IRES project supports the U.S. students’ summer research activities at UoA, where they have the opportunity to engage with leading faculty to gain research experiences, educational enrichment, and participate in innovative exploration centered around convergent research projects on high- and low-power wireless EV charging infrastructure that will enable future EV charging stations and in-motion charging through electrified roadways. In addition to engaging in research, students participate in national experiences while in New Zealand, as well as workforce development, professional development, and academic writing opportunities. These experiences help students understand the societal impact of their research, recognize the value of ethical practices, and understand the history of New Zealand. These activities provide students with experiences that instill an ability to recognize the need for research with societal impact, to learn from and engage with a range of stakeholders to identify societal needs for transportation electrification solutions, and to perform convergent research that considers interdisciplinary, social, and educational aspects of local and global transportation electrification solutions. 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

With the support of the Macromolecular, Supramolecular and Nanochemistry Program in the Division of Chemistry, Professor Christina W. Li of Purdue University will study the defect chemistry of two-dimensional transition metal chalcogenide (TMD) materials. TMDs are unique materials because they can exist stably as a single layer of atoms, which is the thinnest possible form of matter. They can exhibit a broad range of electronic properties depending on the composition and phase of the TMD and can consequently serve as electronic conductors, semiconductors, light absorbers, or catalysts. In addition, the ultra-thin nature of these materials makes their properties highly sensitive to defects on the surface. Under this award, Professor Li’s team will develop strategies to create defects on TMD surfaces in a controlled fashion and to utilize the defects for subsequent functionalization of TMDs to modulate their properties. Understanding the surface chemistry of these ultra-thin materials will guide the development of more efficient electronic and catalytic materials and will have important broader impacts on next-generation semiconductor and energy storage devices. In addition, the team will develop educational demonstrations based on these concepts targeted at K-12 students across Indiana, showcasing energy storage devices and batteries. Under this award, Professor Li's team will study defect-mediated surface functionalization of two-dimensional transition metal dichalcogenide materials. The first goal will be to develop chemical activation strategies to controllably generate specific defect types on the surface of TMDs by utilizing chemical reagents that have strongly nucleophilic, reducing, or basic properties. The characterization methods necessary to study defect density and structure will be developed simultaneously, focusing on reactive titration methods, electronic and vibrational spectroscopy, and high-resolution scanning transmission electron microscopy. Each defect type can subsequently be functionalized with a specific dopant class. In this study, the team will focus on the interaction between chalcogenide vacancies and anionic metal chalcogenide complexes to generate bilayer or multilayer heterostructures. The catalytic and electronic properties of functionalized TMDs will be studied to understand how defect and dopant structure influence TMD band structure and surface chemistry. Finally, the generality of the defect generation and characterization methods will be assessed for a wide range of TMD compositions. 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

Water is a limiting resource for plant production worldwide, and mathematical models indicate that more frequent and longer drought periods will occur. Therefore, utilizing genotypes that produce consistent yields under fluctuating water availability (i.e., yield stability) is critical to ensure food security worldwide. The project team is utilizing maize lines currently used in U.S. maize breeding programs that exhibit a range of yield stability under drought conditions in multi-year field trials. We will use this material to screen for traits that are correlated with variation in yield stability and ultimately identify the genetic bases for yield stability. The information generated in this project will be valuable to public and private institutions that are striving to develop maize lines with high yield stability, as it will identify 1) germplasm that may be important for breeding to improve yield stability, and 2) traits that should be selected for to achieve high yield stability. The project is partnering with Bayer Crop Science, who have committed to support this project through their participation in capacity building. Bayer will provide professional and technical development opportunities for students, staff, and faculty. Reduced and/or erratic water availability will be an issue for most major crop species. Maize is an important system in which to identify and utilize the molecular and genetic mechanisms that underlie drought tolerance and water-use efficiency. Because of the depth of genetic information available, knowledge gained in maize will lead to both a functional application to maize production and a better understanding of other cereal crop species responses to variable environments. Maize germplasm that is currently used in U.S. maize breeding programs have been identified that exhibit diverse levels of yield stability under drought conditions in multi-year field trials. The specific aims of this project are to: 1) estimate a large number of anatomical, morphological, and physiological traits in leaves, stems, and roots of maize that correlate with yield stability under water stress; 2) perform cell-specific transcriptome analyses to characterize molecular components of response systems in roots and leaves, and utilize metabolomic, ionomic, and root microbiota collections to identify specific metabolites, ions, and microbes important to water movement; and 3) validate the cellular and molecular bases for traits identified, including the microbiota, in a screen on selected materials to confirm phenotypic basis for yield stability. This award was funded as part of a lead agency opportunity between NSF, UKRI-BBSRC (U.K. Research and Innovation - Biotechnology and Biological Sciences Research Council; Lead) and DFG (Deutsche Forschungsgemeinschaft/German Research Foundation) where NSF funds the U.S. investigator, UKRI-BBSRC funds the U.K. partner and DFG funds the German partner. 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 Association for Computing Machinery (ACM) MobiCom 2025, will be held in Hong Kong, China on November 4 - 8, 2025. This is the 31st in a series of annual conferences sponsored by ACM SIGMOBILE dedicated to addressing emerging challenges in the core areas of mobile computing and wireless networking. As a top-tier conference, the ACM MobiCom conference attracts significant research contributions that address important research and builds practical working systems, spanning in a wide range of research areas. Moreover, the ACM MobiCom conference features a single-track technical program, where all the accepted high-quality peer-reviewed papers are presented to the entire audience, fostering a sense of community and cohesive experience with interaction and discussion around same core information and common contents. This project supports students from US universities to attend ACM MobiCom 2025 in person. Students will have the opportunity to present their work and be exposed to state-of-the-art developments in the field. They will also have the opportunity to interact with peers from institutions worldwide, meet with senior researchers, and participate in discussions that are likely to shape the future of the field. This grant will target graduate students who will substantially benefit from attending this conference but have limited travel funds. Priority will be given to first-time attendees. 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 objective of this workshop is to identify and articulate research issues and methodologies at the intersection of Artificial Intelligence (AI) and Engineering Design and Systems Engineering (EDSE), with the goal of enabling high societal impact. Recent advances in AI, including generative models, reinforcement learning, digital twins, and human-AI collaboration, are transforming the way engineers design systems, make decisions, and interact with digital tools and data. These advances are reshaping not only the kinds of products and systems that can be designed but also the processes by which they are created and the roles of designers and systems engineers. This workshop will provide a forum for fostering dialog among experts from the EDSE and AI communities, as well as application domains, to examine how AI can amplify EDSE research and practice, and what methodological and theoretical innovations are needed to enable that future. The workshop will build on the outcomes of prior NSF-sponsored EDSE workshops, focusing on the research implications of emerging AI technologies for future engineering design and systems engineering work. Specific emphasis will be placed on identifying new integration mechanisms for AI within EDSE, uncovering high-impact research opportunities, and developing strategies for amplifying collective impact across academia and industry. Expected outcomes include the identification of foundational research questions, the development of methodological frameworks, and the articulation of high-quality research standards to guide the field. The outcomes of the workshop will be disseminated through a workshop report, a curated online video archive, and peer-reviewed journal publications. The workshop will also contribute to the professional development of faculty and researchers across U.S. universities, helping them incorporate cutting-edge insights into their research and educational practices and better prepare engineers for an AI-enabled future. 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.