Georgia Tech Research Corporation

NSF Awards · FY 2024 · 2024-12

With support of the Chemical Synthesis Program in the Division of Chemistry, Justin Kim of the Dana-Farber Cancer Institute is studying chemical processes for the reversible assembly and functionalization of polymers relevant to biology. The processes of interest are designed to be compatible with biological systems (so-called 'bioorthogonal reactions') and are expected to enable the generation of new functionally adaptable biomaterials with many potential applications, including eventually as vehicles for drug delivery and as tissue adhesives for wound closure. The results of this interdisciplinary research are anticipated to lead to important advances in polymer design, synthesis and deployment in biological systems. As part of the broader impacts of the funded project, the PI and other members of the Kim research group will engage in educational outreach activities to help broaden participation in science fields by individuals from groups that have been traditionally underrepresented. A highlight of these efforts will be an interactive workshop for local area high school students that will demonstrate important concepts such as bioorthogonal chemistry (an exciting frontier in science that was recently recognized by the 2022 Nobel Prize in Chemistry) and reversible gelation in materials science. The funded project will explore the use of associative and dissociative bioorthogonal click reactions based on enamine N-oxide motifs with allylic functionalization, to assemble and degrade hydrogels in biologically relevant contexts in a precise stimulus-induced manner. Hydrolytic or enzymatic degradation pathways are commonly used to dissociate polymers or hydrogels from their biological substrates in applications of synthetic biomaterials; however, the reliance on such spontaneous environmentally-driven mechanisms for degradation means that the physical properties of the biomaterial steadily deteriorate over time. This work instead focuses on the synthesis of permanent covalently linked polymer networks whose structural integrity remains intact and uncompromised until such time that removal of the biomaterial is desired: degradation is then triggered by a bioorthogonal chemical reaction. The research explores synthetic methods for accessing enamine N-oxide polymer crosslinkers of varying properties (typically formed by pericyclic group transfer between hydroxylamines and alkynes), strategies for the assembly of step- and chain-growth polymers containing these moieties, and the development of bioorthogonal chemical reactions to reductively cleave the covalent crosslinks. Methods to induce functional changes to the biomaterials of interest through the attachment and removal of various ancillary groups will also be investigated. Efficient chemical reactions that make and break molecular connections in complex biological environments are powerful. As such, the fundamental findings of this work have the potential to impact both academic and industrial science in areas ranging from chemical biology to bioengineering. 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 2024 · 2024-11

The broader impact/commercial potential of this I-Corps project is the development of a novel material implant for the augmentation of bone marrow stimulation for the purpose of improving articular cartilage repair. Initial customer discovery and product-market-fit assessment identified an opportunity within the orthopedic device segment targeting the augmentation of marrow stimulation/microfracture that provides an entry point into the minimally invasive device market. Each year in the United States alone approximately 100,000-160,000 marrow stimulation or microfracture procedures are performed, which involves the removal of damaged cartilage down to subchondral bone followed by micro-drilling to encourage tissue infiltration and cell-guided repair. However, this procedure without augmentation has poor long term outcomes requiring revisions within 2-4 years of the initial surgery. Enhancing minimally invasive repair of these focal defects aims to reduce recovery times, prolong joint health and reduce comorbidities associated with current palliative or restorative therapies. The biomaterial implant offers superior mechanics, handling and fixation compared to existing solutions for marrow augmentation. Beyond the joint space, the material platform technology proposes to utilize both conventional and/or additive manufacturing to improve spatial resolution. Modular manufacturing modalities build a scalable and personalized solution for the future. 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 novel shape memory polymer that allows compression into a small shape that can self-expand into a cartilage defect, allowing for use of an oversized implant to be delivered and expand into the walls of the defect. Polymer synthesis and post-processing parameters have been compiled in a database, allowing for optimization of degradation rates and device mechanics during repair. The chemistry of the polymers that enables 3D printing of the implant allows micron scale spatial control of porosity, pore size, and interconnectivity impacting cell and tissue growth. Preliminary sterilization studies using ethylene oxide to verify compatibility, including cytotoxicity and hemocompatibility studies, have been completed. The materials have also been implanted in mice, rats, and pigs modeling repair of musculoskeletal and cardiovascular defects. This technology proposes to solve the problem of long-term repair outcomes for cartilage defect repair by optimizing the repair mechanics of a Marrow Stimulation Microfracture augmentation procedure. 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 2024 · 2024-11

DMR – 2429163: Engineering Enzyme Vesicles with Stimuli Responsive Partitioning Behavior for High Reactivity and Simultaneous Product Separation Non-technical Summary Enzymes are used to make important products in the pharmaceutical, chemical, food, and detergent industries. Chemical reactions that use enzymes can be performed in water at moderate temperatures, leading to more environmentally friendly processes compared to traditional reactions that use solvents and high temperatures. One difficult challenge is when the starting molecules and final products do not dissolve in water, since enzymes function in water. This project addresses that challenge by using microscopic protein containers, called vesicles. The vesicles can hold the enzymes on their outer surface in the water environment, and simultaneously hold the starting molecules and products at the inner surface surrounded by a liquid-like protein that separates from water. This approach enables both the enzyme and the starting materials to be in their own ideal environment very close to each other so the chemical reaction happens faster than if the vesicles were not used. The vesicles also capture the products after the reaction. Experiments in this project vary the vesicle properties to understand why they improve enzyme reactions and optimize them for maximum reaction. The enzyme chosen for this work is a dehalogenase that can make useful chemicals but can also be used to degrade harmful “forever chemicals” that pollute water in some American communities. Therefore, the results of this research can specifically lead to better processes to degrade forever chemicals or produce valuable chemicals, and the knowledge gained can be applied to a wide variety of enzymes and chemical reactions to make production of important molecules more sustainable in the future. In addition, middle school girls will visit the lab to do hands-on experiments, and local high school students will participate in the vesicle research and professional development and mentoring activities. The proposed plan will provide technical training and critical personal mentoring of graduate and undergraduate students. Results from this work will be shared in the chemical engineering and protein engineering courses and at K-12 school visits and field trips, to inform and inspire students to pursue and persist in STEM fields. Technical Summary This research will create switchable, local hydrophobic environments in an aqueous system for enzymes with hydrophobic substrates to improve reactivity and enable simultaneous reaction and separation. This will be accomplished via creation of enzyme vesicles that are self-assembled from engineered proteins whose hydrophobicity can be turned “on” and “off”, eliminating the need for traditional two-phase solvent-aqueous processes and improving the sustainability of industrial biocatalysis. The vesicle design incorporates enzymes directly on the outer surface in the bulk aqueous phase immediately adjacent to the substrate enriched, hydrophobic inner surface, which increases enzyme conversion of substrate. Dehalogenase enzyme, DmmA, will be used here for proof of concept as it can be applied for both organic synthesis and decontamination of environmental pollutants such as fluorinated compounds that disproportionally impact minority communities, contributing to the broad impact. To achieve enzyme vesicles with controllable substrate partitioning, experimental work will (i) develop and characterize the activity of crosslinked, pH sensitive DmmA vesicles, (ii) assess partitioning of DmmA substrates with different hydrophobicity and molecular weight into vesicles made from protein sequences with varying hydrophobicity, and (iii) demonstrate cyclic DmmA reaction and separation of products by pH switching of vesicles between hydrophobic and hydrophilic states. As intellectual merit, this work will create the design space for enzyme vesicles so that for a given enzyme/substrate/product combination, the best vesicle properties can be selected for optimal aqueous reaction and product separation. In addition, middle school girls will visit the lab to do hands-on experiments, and local high school students will participate in the vesicle research and professional development and mentoring activities. The proposed plan will provide technical training and critical personal mentoring of graduate and undergraduate students. Results from this work will be shared in the chemical engineering and protein engineering courses and at K-12 school visits and field trips, to inform and inspire students to pursue and persist in STEM fields. 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 2024 · 2024-10

This IUSE Level 2 Engaged Student Learning project aims to serve the national interest by increasing opportunities for training and enhancing safety in the construction industry, The industry is gearing towards wearable robots as an ergonomic intervention that enhances human strength and performance while reducing muscle fatigue and stress. Consequently, it is crucial to equip construction engineering and management (CEM) students with the skills needed to implement this intervention effectively. This research aims to investigate an interactive virtual reality environment designed to develop CEM students' competencies in implementing wearable robot solutions for addressing ergonomic risks. The proposed research will potentially benefit the construction industry by providing the future workforce with the competencies to advance wearable robot solutions. This interdisciplinary research proposes to equip CEM students with relevant human-wearable robot interaction competencies through guided experiential learning in an interactive virtual reality environment called ViRLE. A first research goal is to identify the required skills, knowledge, and abilities for deploying wearable robot solutions by interviewing industry practitioners. Next, the project involves plans to integrate tools from wearable robots, virtual reality, tangible interactions, and sensing technologies to develop ViRLE. The project team intends to apply constructivism and cognitive apprenticeship theories to design activities and study how use of the technology impacts student learning. The research team then will attempt to identify the characteristics of ViRLE that facilitate interaction with wearable robots in reducing ergonomic exposures of construction tasks. Finally, planned implementation of ViRLE will occur in two institutions with diverse student populations to assess its potential to support the acquisition of these competencies. The research team envisages providing an interactive learning environment that is suitable for students of diverse demographics. 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 2024 · 2024-10

Conventional nitrogen removal in wastewater treatment systems is highly energy intensive, accounting for nearly 30% of the total energy demand in the plant. In addition, conventional nitrogen removal processes emit nitrous oxide to the atmosphere, which is a highly potent greenhouse gas. Thus, the energy demand and greenhouse gas emissions of nitrogen removal processes have a large impact on the carbon footprint of wastewater treatment. This project will advance novel biological treatment strategies that use highly specialized comammox and anammox bacteria for nitrogen removal as opposed to conventional nitrifying bacteria. This process holds great promise to reduce the carbon footprint of wastewater treatment by lowering energy demand and reducing greenhouse gas emissions. This research project also includes substantial collaboration with wastewater utilities ensuring that the outcomes will rapidly translate to industry benefits. This collaboration also provides a unique opportunity for training of undergraduate and graduate students by academic researchers and industry practitioners. The goal of this project is to advance the sustainability of wastewater treatment by systematically investigating the recently discovered synergistic nitrogen removal by comammox and anammox bacteria in a full-scale mainstream nitrogen removal process. Specific research objectives designed to achieve this goal include: i) evaluating the impact of nitrogen loading rates on the synergy between comammox and anammox bacteria, ii) determining the impact of dissolved oxygen concentrations on mechanisms governing nitrite provision to anammox bacteria, and iii) conducting pilot-scale testing of two different process configurations that leverage comammox and anammox bacterial synergy for nitrogen removal. Using anammox bacteria in mainstream nitrogen removal has the potential to reduce aeration costs by nearly 60%, thus significantly reducing the carbon footprint of the wastewater treatment sector. Further, shifting aerobic nitrification from strict ammonia-oxidizing bacteria to comammox bacteria combined with potentially eliminating the need for denitrification will substantially reduce nitrous oxide emissions from wastewater treatment. A better understanding of comammox-anammox synergy will also have a beneficial impact on nitrogen pollution management in other sectors like agriculture where managing the nitrogen cycle is becoming increasingly urgent. This industry-facing interdisciplinary project will also provide a unique opportunity for graduate and undergraduate researchers to be mentored by academics and industry practitioners. These training and stakeholder engagement efforts will contribute to a skilled and diverse STEM workforce. 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 2024 · 2024-10

In today's Big Data Era, the relentless exponential increase of data generation, especially of real-time data from personal daily activities coupled with emerging applications, not only offers great and unprecedented opportunities, but also imposes a significant challenge to process and transmit the ever-increasing large volume and varieties of data in timely manner and avoid being drown in the constantly fast-expanding gigantic data sea. One of the key enablers to achieve this goal is an energy-efficient and ultra-high-data-rate wireless communication system that matches and, at the same time, scales with the data generation rate. Moreover, wireless communication systems are vulnerable to data intrusion with the increasing number of access networks and nodes in dynamic and open communication environments. This forms a big challenge to wireless cybersecurity. Therefore, to satisfy the needs in the Big Data Era, the next generation wireless communication systems with improved energy efficiency and ultra-high data rate while achieving enhanced security is demanded. The research will develop a new reconfigurable and scalable transceiver phased array system, operating at mm-wave to sub-mm-wave frequencies, to achieve these three objectives. To overcome the many performance challenges at such high frequencies, this project will develop several key enabling and new techniques at different design levels, including system configuration, transceiver architecture, and circuit design. The success of this research will advance scientific understanding and create a new design methodology to achieve unparalleled data rates and energy efficiency, which will broadly impact the wireless industry and benefit the society. Moreover, this project will train future engineers and scientists for the fast-growing data-driven industries, with special efforts to promote diversity by training more female and minority students. The project will develop a reconfigurable and scalable wireless communication system, operating at mm-wave to sub-mm-wave frequencies, that can be efficiently reconfigured into three operation modes: ultra-high data rate for short distance, high data rate for medium distance, and medium data rate for long distance. The array architecture is based on coupled oscillators for high-efficient frequency tuning and beam forming. The unique tuning scheme allows the array size to be scaled effectively for different operation modes to be deployed in different application scenarios without redesigning the whole system. The new direct antenna modulation scheme enables ultra-high data rates and boosts transmitter energy efficiency by mitigating conventional antenna bandwidth constraints and eliminating linear power amplifiers. And the proposed high gain and low noise mixer structure extracts signal phase information to enable high-order demodulation scheme with enhanced receiver noise and gain performance and reduced power consumption. In addition, the proposed redundancy mapping scheme offers secure wireless communications without extra power and communication overheads. If successful, the system's data rate and energy efficiency will be orders of magnitude higher than existing technologies and therefore the new system will open a new door for secure and ultra-high-speed wireless applications. This project will also investigate the design methodologies on how to achieve the highest frequency/speed with the best energy efficiency systematically, from system and circuit levels down to device level. The transformative design methodologies are expected to benefit other wireless applications, such as radar, imaging, and sensing. 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 2024 · 2024-10

Personal computers and mobile phones have been the primary platforms for interaction with the digital world, pervading every corner of the modern society in the past few decades. Augmented Reality/Virtual Reality (AR/VR) is regarded as the general human-oriented computing platform for the next decade. In the smart headset, eye tracking provides crucial information on human perception, and it is a key pillar that supports many AR/VR applications such as displays, user interfaces, and foveated rendering. Hence, real-time processing of eye tracking workloads is critical, but modern eye tracking algorithms are much slower than the desired specifications (over an order of magnitude slower). Eye movements including saccade dynamics are one of the fastest movements in human body. Therefore, to accurately track eye movements, we need to perform image segmentation tasks at a very high frame rate (>1000 frames per second, FPS). However, this involves a large amount of data movement between the image sensor and the processor that performs the segmentation tasks. In the current practice, integration of frontend complementary-metal-oxide-semiconductor (CMOS) image sensor and backend microprocessor is slow and inefficient as they are generally placed in separate packages. To efficiently perform high frame rate processing, in-pixel and near-pixel compute paradigms have been proposed to reduce the latency and energy consumption from the costly data movements to provide >100× reduction in communication latency by placing compute close to the source of data. The outcome of this award will have synergies with other national efforts to revamp domestic semiconductor research and development under the CHIPS and Science Act. Besides the AR/VR applications, tracking fast eye movements could be useful in biomedical applications such as early diagnosis of Alzheimer’s and Parkinson’s diseases. The objective of the research and education integration is to train the next generation of workforce with domain expertise and interdisciplinary skills in the broad area of semiconductor device, integrated circuit and advanced packaging. This award aims to advance the software-hardware co-design for in-pixel and near-pixel compute for eye tracking in AR/VR applications. A multi-mode image sensor that supports eye tracking is proposed, featuring a successive frame differencing method for the event map generation with in-pixel compute chiplet. It will be capable of providing both an event map at a higher frame rate as well as a full resolution image at a lower frame rate. Moreover, the near-pixel compute chiplet will run eye tracking inference with dual-mode deep neural networks for both high accuracy and high frame rate. The proposed research activities also include exploring the fundamental device technologies such as characterizing the amorphous oxide semiconductor transistor for pixel read-out circuitry and exploring the advanced packaging techniques for system-level heterogeneous integration. Silicon tape-outs are planned to validate the in-pixel and near-pixel compute chiplets with a pathway for heterogeneous integration on a silicon interposer. 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 2024 · 2024-10

This project aims to develop software systems for robust and efficient support of interactive Artificial Intelligence-enabled services. An increasing number of applications rely on pre-trained Machine Learning models for their operation, such as computer vision tasks, natural language processing, and recommendation systems. Cyber-physical systems (e.g., autonomous navigation) also operate in a tight control loop of sensing the environment, perception, and performing corresponding actions.These applications collectively can be characterized by requiring Machine Learning models to be both accurate and able to respond within strict deadlines. Such responsiveness is mission-critical, as a late response can bear undesirable consequences, no matter how accurate. And yet, these applications are deployed in increasingly heterogeneous and unpredictable environments. This project’s novelties lie in building software frameworks that are able to navigate this accuracy/responsiveness tradeoff space automatically with high efficiency in real time. Its broader significance and importance lies in enabling robust AI-powered applications to withstand unpredictable deployment conditions in domains spanning healthcare, cyber-physical systems, online recommendation systems, and AI-assisted search and rescue operations. To address these challenges, this project develops agile mechanisms and policies for serving a family of models instead of a fixed model for a single prediction task. Its ability to activate different models near-instantaneously in-situ enables operating across heterogeneous devices without retraining the neural network for the served model. This saves on training cost – amortized over multiple heterogeneous devices and per-device dynamic deployment conditions (e.g., latency budget variation, processor frequency scaling, battery level, etc.) These unpredictable sources of variability are supported by the novel combination of this project’s mechanism and policy that schedules a stream of ML prediction tasks. The project generalizes inference serving, decoupling ML serving mechanisms from the policies that control query scheduling decisions. This project aims to produce system software artifacts that will be open sourced for the general public. Research outcomes from this project will be incorporated in several courses at both graduate and undergraduate level, collectively attracting more than 165 students annually. This is in addition to the prompt dissemination of research outcomes to the public through conference proceedings, seminar talks, and invited keynotes. 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 2024 · 2024-10

Cloud-based data stores employ data replication combined with algorithms that allow consistent, concurrent access to data to offer low response times to their global client-bases. Erasure coding (EC) is a generalization of replication known to be more storage efficient for the same fault tolerance in theory. However, reaping the benefits of EC for geo-distributed data stores in practice poses several new scientific challenges. The overarching goal of this project is to address these challenges by developing a comprehensive understanding of the response time vs. cost trade-off for EC-based consistent geo-distributed stores. The research will be conducted in the following four thrusts: (i) development of EC techniques and distributed algorithms to satisfy consistency criteria known as eventual and causal consistency, (ii) development of efficient, EC-compatible reconfiguration algorithms that provably maintain consistency despite changes in object configurations, (ii) development of analytical models of performance and cost, and associated optimization frameworks, and (iv) integration of developed techniques into Apache Cassandra. The proposed research will be conducted in the context of the public cloud, where inter-DC latencies and pricing information are readily available allowing competing schemes to be compared in a fair manner. The proposed research could lead to cost reductions for geo-distributed data stores hosted on public clouds, which are fundamental building blocks for several applications including collaborative editing, social media, financial transactions, reservation systems, and multi-player gaming. The validation plan involves experiments performed using an Apache Cassandra-based prototype that will be made open source. The prototype will serve as a testbed for other researchers and engineers to plug in their own algorithms and compare findings. The proposed research has a cross-disciplinary nature and will be supplemented with an education plan that involves development of survey articles and curricular integration. 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 2024 · 2024-10

This project explores the design and development of acoustic technology as a complementary modality to radio frequency (RF) for indoor smart Internet of Things (IoT) applications. However, today’s widely available smart devices with in-built acoustic interfaces are neither cost-effective nor energy-efficient, while acoustic tags designed for low-power applications have exposed significantly limited performance in one or more dimensions of throughput, range and energy efficiency. Towards enabling a myriad of practical IoT applications with acoustics, the proposed research explores a novel collaborative paradigm, where it brings together multiple, distributed low-power acoustic tags to enable on-the-fly transmit time-delay beamforming (2D and 3D) to a target receiver, thereby delivering significant performance gains in range, throughput and energy efficiency simultaneously. It can impact our society in multiple ways -- from novel IoT applications to more broadly enabling low-power acoustic communication in a wide array of application domains ranging from secure acoustic spaces to immersive media and underwater sensing. It aims to inform relevant wireless and acoustic professional organizations by demonstrating the potential of such an energy-efficient distributed acoustic paradigm. It will also impact the audio industry that is actively working on speaker arrays for targeted, immersive audio in numerous consumer applications. The proposed research aims to take an important step towards this vision by addressing three key challenges: (i) scalability -- enabling an on-demand beamforming array that can implicitly synchronize and share data amongst a distributed set of tags and scale its performance with the array size over a wideband; (ii) coexistence -- enabling such wideband performance without causing or being impacted by audible interference created in the process; and (iii) multiple access -- efficiently allowing multiple tags to access the channel in creating practical IoT applications that often include both near and far-field deployments. Through the design of an analytically-sound collaborative beamforming framework, a novel distributed acoustic tag-array is proposed that brings together multiple low-power individual tags on-demand through intelligent distributed mechanisms for time-delay (2D and 3D) beamforming and angle-dependent signal whitening. The latter enable wideband operations in both near and far-field, while suppressing interference to unwanted regions and allowing multiple such tag-arrays to coexist. The proposed paradigm is expected to be implemented and tested rigorously in real world, while enabling two practical acoustic IoT applications in inventory management and targeted audio delivery. The project paves the way for future IoT systems to leverage acoustic interfaces with scalable performance as a complementary modality to RF. 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 2024 · 2024-10

This workshop will convene experts from academia, industry, and government to explore the transformative potential of generative Artificial Intelligence (AI) in hardware design. The workshop will be co-located with the 57th International Symposium on Microarchitecture (MICRO), sponsored by the Association for Computing Machinery and the Institute of Electrical and Electronics Engineers, a premier forum for disseminating research in the fields of architecture, compilers, chips, and systems. Scheduled as a one-day event, it will feature keynote speeches, invited talks, a panel, and roundtable discussions, focusing on how generative AI can revolutionize hardware design and how advancements in hardware can enhance AI capabilities. The event will bring together researchers with diverse backgrounds to ensure a rich and inclusive exchange of ideas. It aims to foster collaborative research across disciplines, identify new research opportunities, and address challenges at the intersection of AI and hardware. This initiative supports the NSF's mission to promote the progress of science, advance national health, prosperity, and welfare, and secure the national defense. Generative AI has the potential to significantly accelerate the design and optimization of computer architectures, enabling rapid prototyping and verification of innovative hardware solutions. Despite its potential, the development of AI-integrated hardware solutions faces substantial challenges due to the need for interdisciplinary knowledge spanning machine learning, systems engineering, and computer architecture. This workshop will address these challenges by bringing together leading experts to share insights, foster innovation, and discuss cutting-edge research. The workshop will highlight the reciprocal impact of generative AI and hardware design, where advancements in one area significantly enhance the other. Leveraging recent breakthroughs in generative AI, the workshop will explore how innovative hardware design can subsequently advance AI capabilities, promoting faster and more energy-efficient hardware systems that intuitively adapt to AI demands. 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 2024 · 2024-10

This award funds a workshop on fractured rock masses, a rapidly emerging topic that plays a crucial role in many pressing energy and environmental challenges, such as underground construction, natural hazards, geothermal energy, mining, environmental restoration, carbon sequestration, nuclear waste geological storage, and subsurface energy storage. Developing innovative solutions to these challenges relies on a deep understanding of fractured rock masses, which however is hampered by the inherent complexities that arise due to coupled multi-physics processes and multi-scale fracture-matrix interaction effects. Studying fractured rock masses promotes multiple scientific fields of geotechnical engineering, geosciences, physics, sensors and sensing, seismology, mining and minerals, subsurface resources, and renewable energy. This workshop will reach beyond the associated engineering and science communities, including improved understanding of scientific problems, augmented capabilities in scientific tools and infrastructure, deepened insight into innovative engineering solutions, cross-education and collaborations among researchers, and a sustained pipeline of workforce and STEM education. The workshop will also stimulate interest among young investigators who are future leaders in the field and cultivate a collaborative and innovative community in addressing energy security, environmental issues, and societal challenges. This workshop will convene approximately 30 individuals with diverse backgrounds and demographics to distill the most important science questions related to fractured rock masses, develop strategies to address the challenges, conceive unique and unprecedented infrastructure and facilities, and foster a collaborative and educational community to educate the next generation of engineers working on fractured rock masses. The workshop will create the opportunity for cross-disciplinary experts to discuss topics that reside at the interface of multiple fields such as self-similar fracture topology, the cubic law of transmissivity, complexities in fractured rock mechanics, and the availability of robust engineering analysis and design tools. Key observations and conclusions from this workshop will be summarized and disseminated through a white paper. To attain the workshop objectives, a series of activities before, during, and after the workshop will be organized to maximize the productivity of the gathering and ensure the broad distribution of the workshop outcomes. 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 2024 · 2024-10

This three-year renewal RET Site: Collaborative Research: Research Experiences for Teachers across the National Nanotechnology Coordinated Infrastructure is hosted by the Georgia Institute of Technology, the University of Minnesota, and the University of Nebraska, Lincoln. Nanoscale science and engineering is interdisciplinary and cuts across all science and engineering disciplines. As part of the National Nanotechnology Coordinated Infrastructure (NNCI) this program supports 12 in-service, high school and community college faculty each year. Participants will engage in high-quality, nanoscale science and engineering (NSE) hands-on research in state-of-the-art nanotechnology facilities at NNCI sites for 6 weeks during the summer. Educators will complete a hands-on research project in NSE during the summer with continuing support during the academic year. This RET program spanning three NNCI sites allows participants access to a wider variety of NSE research than would be available at a single-site and exposes participants to the NSE needs of industry and related career opportunities across the nation. Project activities will strengthen participants’ knowledge and understanding about broad educational, industrial, and societal NSE activities and how to motivate their students to explore STEM and NSE fields that may lead and provide them with satisfying and lifelong STEM careers. This three-year renewal RET Site: Collaborative Research: Research Experiences for Teachers across the National Nanotechnology Coordinated Infrastructure (NNCI) is hosted by the Georgia Institute of Technology, the University of Minnesota, and the University of Nebraska, Lincoln. Nanoscale science and engineering is interdisciplinary and cuts across all science and engineering disciplines. The program offers a wide array of topics such as flexible electronics, nanomotors, batteries, environmental filtration, and medical diagnosis of diseases. With support from faculty, mentors, and RET coordinators, the RETs will develop curriculum materials to bring their NSE research back to their classrooms. During the academic year, faculty and mentors will visit the RET classrooms to assist with the implementation and further development of the curriculum modules. This RET program spanning three NNCI sites allows participants access to a wider variety of NSE research than would be available at a single-site and exposes participants to the NSE needs of industry and related career opportunities across the nation. The objectives are to grow a multi-site cohort of educators with research experiences that reflect broad educational, industrial, and societal NSE activities; build and disseminate a library of NSE educational materials; highlight the work of NNCI cohort by attending each sites state science teaching association annual meeting; and encourage RETs to present at professional society meetings. Webinars will be held across all participating NNCI sites to enable teachers to learn about NSE industries and careers as well as discuss their modules. The RET program promotes networking opportunities through participation in on-line presentations and webinars, a Slack group, the yearly state science teacher conferences, professional society conferences, and an in-person convocation. This project is partially supported by the Division of Electrical, Communications, and Cyber Systems, the Established Program to Stimulate Competitive Research (EPSCOR), and the Division of Engineering Education and Research Centers. 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 2024 · 2024-10

Understanding and predicting earthquakes is a critical endeavor that has profound implications for public safety and disaster preparedness. By developing cutting-edge machine learning models and algorithms, this project seeks to uncover the intricate dynamics of earthquakes, potentially identifying precursory signals that precede major seismic events. The broader impact of this work includes enhancing our ability to forecast earthquakes more accurately, thus mitigating risks and improving resilience in communities prone to seismic activities. Additionally, the project will create open-source tools accessible to researchers and practitioners worldwide, fostering global collaboration and knowledge sharing. This initiative not only aims to advance scientific understanding but also to inspire and educate the next generation of geoscientists through engaging sonification/animation products and educational programs. This research addresses the challenge of understanding earthquake physics and forecasting its collective behaviors by utilizing high-resolution, high-dimensional earthquake catalogs and continuous geophysical measurements. Traditional forecasting models struggle with the complexity of such detailed data; thus, this project proposes novel approaches grounded in marked temporal point processes. The key strategies include developing advanced Hawkes process models with deep neural triggering kernels to gain nuanced insights into earthquake dynamics, creating a novel generative framework to explore complex seismic patterns, and applying these methods to recent earthquake sequences in California, Japan, and Türkiye. The project will produce open-source software tools to support these efforts. The intellectual merit lies in integrating advanced statistical models, machine learning techniques, and high-resolution earthquake catalogs to address longstanding challenges in geoscience. By enhancing the representation of earthquake dynamics with deep neural triggering kernels within Hawkes process models, the project aims to overcome limitations of traditional forecasting methods. The generative framework for marked temporal point processes will enable systematic exploration of intricate seismic patterns. International collaborations and the development of accessible, open-source resources exemplify a commitment to impactful and practical research. Additionally, the project will offer Research Experiences for Undergraduates (REU) at Georgia Tech, promoting interdisciplinary collaboration and broadening participation in geosciences. The collaboration between an early career machine learning PI and a mid-career earthquake seismologist further underscores the project's innovative and interdisciplinary nature. 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 2024 · 2024-10

Understanding social animal behaviors has become a major focus in computational neuroscience due to its ability to reveal complex brain-behavior connections. Despite advances in studying animal behavior, challenges remain in comprehending natural and social behaviors over long periods. Imagine the behavior of two mice moving around in an open field over several days. Many behaviors and interactions can happen over shorter and longer time scales. In just a few seconds or minutes, a mouse might groom itself or dart around, or two mice might sniff each other. Over hours or days, they might look for water or food together, or one might grab food from the other. Each behavior reflects different cognitive processes and has potential to connect with neural functions. The goal of behavior analysis is to extract and parse such complicated multi-animal behavior from days of video recordings, accounting for multiple decision-making processes and actions involving multiple individuals. This project seeks to provide accurate and reliable machine learning methods for identifying short-term actions, understanding long-term decision-making, and analyzing multi-animal interactions. These tools will offer new insights into natural and social animal behaviors and enhance the understanding of the relationships among brain regions, neural functions, and behaviors, especially in naturalistic and real-world contexts. This project introduces advanced computational methods designed to analyze complex multi-animal behaviors. Aim 1 focuses on developing new self-supervised learning techniques for behavior segmentation, breaking down long behavior sequences into short, distinct actions like grooming or sniffing. Aim 2 addresses animal behavior as a series of decisions influenced by goals and rewards. A novel inverse reinforcement learning framework will be developed to explore how animals make decisions and how these decisions change over time. Aim 3 introduces a new spatiotemporal graph to capture interactions between animals and to study short-term actions and long-term decision-making in social contexts. These aims are interconnected to understand how behaviors are generated across different time scales and spatial contexts. The ultimate goal is to create a practical set of accurate, interpretable, and reliable machine learning models for analyzing multi-animal behaviors. Insights gained will reveal neuro-behavioral correlations across brain regions and functions, potentially impacting fields such as AI, robotics, cognitive science, and psychology. 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 2024 · 2024-10

This Research Experiences in STEM Settings (RESS) project aims to serve the national need of preparing highly qualified secondary STEM teachers equipped with authentic research experiences. Additionally, this project will support 30 pre-service teachers in STEM fields by offering six-week research experiences in university labs and opportunities to develop lesson plans based on their research experience. The proposed project components will enable high-achieving prospective teachers to become secondary STEM teachers with extensive expertise in STEM content, research practices, and teaching skills. This project at Georgia Institute of Technology includes partnerships with Clayton State University, Mercer University, and local school districts. Project goals include developing a sustainable STEM Research Experience for Pre-Service Teachers (REPST) summer program and enhancing participants’ content knowledge and pedagogical skills through authentic research and mentoring experiences. Over three years, the project will engage 30 STEM undergraduate and master’s level pre-service teachers in these research experiences. This project will be iteratively evaluated. Evaluation of the project will be guided by the following evaluation questions: (a) How do authentic research experiences impact pre-service teachers’ STEM content knowledge and teaching practices? (b) How do these experiences influence participants’ ability to transfer research practices to their teaching? The results of this project will be disseminated to help enhance the field. This RESS project is supported through the Robert Noyce Teacher Scholarship Program (Noyce). The Noyce program supports talented STEM undergraduate majors and professionals to become effective K-12 STEM teachers and experienced, exemplary K-12 teachers to become STEM master teachers in high-need school districts. It also supports research on the effectiveness and retention of K-12 STEM teachers in high-need school districts. 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 2024 · 2024-10

This project aims to broaden access to large visual learning models for small businesses, developers and users with limited computational, data, or expert resources. Such large learned models have the potential to revolutionize multiple application areas from manufacturing to sustainability solutions to healthcare. While technology is emerging in relevance and commercial viability, it remains expensive to deploy successfully and hence its usefulness is often limited to companies and entities with maximal resources. The research from this project will focus on democratizing this technology by reducing the amount of data, computation, and human expertise needed to create and deploy application specific models. The fundamental barriers to broadened access are computationally intensive inference requirements, assumption of significant data access, and poor model interpretability. The project aims to address these challenges through an integrated plan that proposes novel learning approaches for fast model specialization with limited data, advances in inference algorithms to reduce redundancy and model reuse to increase inference and training efficiency, and corresponding advancements in visualization and deployment tools to aid usability. The project team will work together with under-resourced partners to assess the effectiveness of the new approach. While the study will focus first on computer vision models, it is expected that these advances will translate to other modalities with transformer-based models such as natural language processing and speech processing. In addition to deployment of technological advances with potential users, the findings from this project will also be integrated into undergraduate courses and research opportunities. 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 2024 · 2024-10

Learning-enabled systems have been rapidly increasing in size, acquiring new capabilities. These systems are typically deployed in complex operating environments, so their safety is extremely important. Ensuring safety requires that systems are robust to extreme events while we can monitor them for anomalous and unsafe behavior. While traditional machine learning systems are evaluated pointwise with respect to a fixed test set, such static coverage provides only limited assurance when exposed to unprecedented conditions in complex operating environments. One key question that still remains unanswered is: How can we design and deploy learning-enabled systems that can be robust to extreme events while monitoring them for anomalous and unsafe behavior by synthesizing model-free and model-based techniques? The overarching goal of the proposed research is to establish a framework that leads to the design and implementation of learning-enabled systems in which safety is ensured with high levels of confidence. The framework will leverage tools from control theory, multi-agent autonomy, and formal methods for rigorously probabilistic reasoning to yield safe learning-enabled systems. The expected outcome of this project will yield safe model-free, mode-based, and interacting model-free and model-based learning-enabled systems with sound design principles that practitioners could leverage to achieve safety specifications. The proposed research could effectively facilitate safe learning-enabled systems even within complex environments while monitoring them for anomalous and unsafe behavior. It will yield learning-enabled systems that could be deployed in complex operating environments while ensuring that the systems will be robust to extreme events and monitoring them for anomalous and unsafe behavior. The fundamental knowledge created in the proposed research will be the basis upon which future-safe autonomous systems with embodied intelligence can be built. 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 2024 · 2024-10

This Faculty Early Career Development Program (CAREER) grant will fund research that clarifies how loss of integrity of bolted joints affects the resilience and progression to failure of safety- and reliability-critical mechanical structures, with application to vehicles, industrial equipment, biomedical implants, space telescopes, and playground equipment, thereby promoting the progress of science, and advancing the national prosperity and welfare. Loose bolts and screws are a common problem in US infrastructure. Structural failures often have catastrophic consequences, for example, resulting in vehicle crashes or train derailments that lead to casualties, economic loss, and environmental damage. Little is known about how a structure’s dynamics influences the loosening of bolted joints over long periods of operation. Models calibrated using data from the time of assembly fail to enable optimized preventative maintenance over a structure’s lifetime. This project overcomes these challenges by developing predictive modeling approaches that capture the coupling between structural dynamics and the integrity of bolted joints over timespans consistent with the structure’s service life, leading to improved health monitoring of aging infrastructure, toughened designs against the impacts of extreme weather, and more reliable energy generation from renewable sources. The project advances STEM learning and enhances diversity through a “Teach for Discovery” approach that engages students’ natural curiosity through game-based learning and virtual reality experiences. Advise from, and outreach to, industry and national labs ensures that the research activities are informed by industry needs and that results are accessible and effectively assessed. This research aims to develop the foundations of a modeling and validation framework for predicting the long-term nonlinear dynamics of assembled structures with loosening bolted joints. It achieves this aim partly by combining new experimental and modeling techniques to map the strains around a tightened bolt head to the underlying contact conditions inside the interface as the joint evolves dynamically, for example, to test the hypothesis that the rate of loosening is independent of the relative displacement across the interface. Models relating bolt tension to joint stiffness and damping and differential equations governing the evolution of the tensions are then derived from experimental measurements. Finally, the theory of nonlinear normal modes is expanded to incorporate bolt tension as a dependent variable and used to characterize the influence of internal resonances on the progression of local damage. 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 2024 · 2024-10

Dielectric constant or permittivity is an important property of materials and biological cells. The ubiquitous and fast sensing of them not only will help us better understand ourselves and surrounding environments, but can also help prevent devastating events in society, such as global pandemics, by timely detection. Ubiquitous dielectric sensing requires not only high performances such as accurate, robust, and trustworthy results, but also fast response, small form factor, low cost, low power consumption, etc., for sensors to be widely deployed in different scenarios and applications. Existing optical and electronic dielectric sensing techniques face great challenges to meet all the requirements. Therefore, this project aims to bridge the critical gap between optical sensing and electronic sensing and develop ubiquitous dielectric sensors meeting all the above requirements for daily usage. The research results of the proposed project are expected to not only directly advance ubiquitous dielectric sensing for human daily lives, but also impact the societies at large. The successfully demonstrated sub-THz/THz design techniques proposed in this project will advance knowledge and facilitate exploration of this underutilized spectrum. In the educational front, the principal investigator (PI) will broadly disseminate the research results through presentations and publications in scientific conferences and journals, as well as integrate research with education and outreach programs. The PI is committed to engaging and retaining students from underrepresented groups in engineering and STEM fields and attracting minority and female students through various local programs. The PI will continue her conscientious outreach efforts to local high schools to inspire K-12 students, especially minorities and socioeconomically disadvantaged students, to join the engineering world. To materialize the overarching goal of ubiquitous access to dielectric sensors for daily lives, the proposed research will investigate fast, accurate, compact, trustable, low cost and power (FACTCoP) sub-THz/THz ring-resonator-based dielectric sensors to leverage the benefits of both high-performance optical micro-ring resonator sensors and unparalleled on-chip signal processing with THz speed from advanced semiconductor devices and circuits. It integrates three coherent major tasks enabled by new design ideas and schemes. The first task is to boost sensitivity by intensifying evanescent electromagnetic fields for enhanced wave-matter interactions through multi-dimensional sub-THz/THz slot rings and waveguides and using phase-based sensing modality. The second task is to develop a holistic noise suppression scheme to significantly reduce various noise sources, including transmitter signal phase noise, receiver flicker noise, common-mode noise, ambient noise coupling, as well as broadband thermal noise. The systematic noise suppression scheme is to improve minimum detectable signal for enhanced sensor resolution. The third task is to develop low-power, low-noise integrated signal generator in transmitter and signal detection and processing in receiver by exploring new innovative design ideas and techniques for sub-THz/THz integrated circuits and systems, such as a sub-THz/THz sub-sampling phase-locked loop on the transmitter side, multi-path noise cancellation on the receiver side. In addition to the sensor design and development, the PI and her team will also address the following key questions in their research: 1) what are the ultimate noise constraints for sub-THz/THz circuits in different noise domains? 2) with the proposed noise suppression scheme, what is the theoretically achievable sensing resolution? 3) under real system hardware implementation constraints, such as mismatches and parasitics in circuits and components, what are the practically achievable sensing resolutions? 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 2024 · 2024-10

Distributed computing plays a central role in enabling modern machine learning and artificial intelligence applications. This project develops new theoretical frameworks, analyses, and techniques for distributed computing and machine learning aiming to (i) accelerate the computation time by overcoming system bottlenecks, (ii) ensure accurate computation in the presence of hardware errors and faults, and (iii) enable data-processing approaches that adhere to data-privacy constraints. These attributes will be enabled by inducing a controlled amount of redundancy in computations based on coding theory - a field that has enabled modern data communication and storage technologies. To obtain a realistic understanding of what is possible in the long run, the developed techniques will be accompanied by fundamental bounds on the tradeoffs between computation accuracy, data privacy, error tolerance, and redundancy overheads. The project will disseminate outcomes and enable awareness of developed research to the broader scientific community through publications, tutorials, and curricular integration. Coded computing is a sub-area of information and coding theory that induces redundancy into distributed computing. Coded computing has emerged as a promising paradigm to relieve straggler, communication, and data-privacy bottlenecks in large-scale distributed machine learning. Yet, state-of-the-art coded-computing techniques, mostly devised to enable exact reconstruction of the computation output, have fundamental efficiency limitations, particularly for nonlinear computation tasks. This project develops techniques for approximate coded computing, wherein the decoder aims to obtain the function output within a prescribed distortion limit, and data-privacy constraints are posed as limits on differential-privacy parameters. The research will be conducted in three closely connected thrusts: (i) coding schemes for fault-tolerant approximate matrix multiplication, (ii) coding schemes for fault-tolerant approximate nonlinear computations beyond matrix multiplications, and (iii) coding schemes for differentially private computations. The techniques developed will combine ideas from information and coding theories, mathematical approximation theory, and differential privacy. 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 2024 · 2024-09

The key aim of this project is to examine the role of evidence-informed science diplomacy as a strategic instrument to address opportunities for strengthening and threats to maintaining democracy, governance, and trust (DGT). This work contributes towards a conceptualization of a concept with rising importance on the international stage and presents an innovative approach to integrating metascientific evidence into the practice of diplomacy. The intended impact of the project is to strengthen DGT by supporting data gathering and production of inclusive indicators and amplifying trust in science within and across nations. This research will contribute to our theoretical, historical, and empirical understanding of science diplomacy and advance knowledge across several fields (e.g., science and technology policy, science communication, political science, and international relations). This project will utilize qualitative and quantitative methods to (1) understand the relationship between science diplomacy and DGT; (2) conceptualize and operationalize metascience observatories and investigate the extent to which they can be leveraged to improve science diplomacy; and (3) explore how threats to DGT could be mitigated and opportunities seized through inclusive metascience observatories. The outputs will include both academic-oriented products, as well as communications to policymakers and the wider public. In addition to these products, outcomes will include communities of practice and training opportunities. The long-term goal of the project is to strengthen DGT through evidenced-informed science diplomacy. 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 2024 · 2024-09

Computational quantum chemistry provides accurate descriptions of molecules, and it has become a standard research tool in chemistry, biochemistry, chemical engineering, materials science, and other fields. However, the equations in quantum chemistry are very complicated. This means that they are very hard to convert into computer programs, and also that those computer programs can take a very long time to run. The long computation times mean that many theoretical chemists are actively engaged in developing new theoretical models that yield faster computations, while hopefully having a minimal impact on accuracy. However, these new methods also tend to involve complicated equations that are hard to implement in computer programs. Because the programs are very complicated, they are difficult to adapt to new computer hardware that could make them run faster. This collaborative project will develop a software framework to make it much easier to implement advanced quantum chemistry methods on emerging hardware. It involves a team of experts in quantum chemistry and computer science from Georgia Tech, the University of Georgia, Virginia Tech, and the University of Memphis. This team will develop a library to efficiently handle the matrices and tensors that appear in quantum chemistry equations, and to make it easy for programmers to implement quantum chemistry methods by writing code that looks more like the equations. It will also develop a library to compute the electron repulsion integrals that are central to quantum chemistry on graphics processing units. These tools will be thoroughly tested by using them to implement several advanced theoretical methods, including coupled-cluster, relativistic, and real-time electron dynamics methods. These implementations will test the libraries and will also provide advanced simulation techniques to researchers. This award by the NSF Office of Advanced Cyberinfrastructure is jointly supported by the Division of Chemistry. 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 2024 · 2024-09

Many real-life decision-making tasks involve working with uncertainty, where decisions must be made to minimize costs affected by underlying factors that influence uncertainty. For example, in healthcare, allocating resources for diseases require anticipating future cases in various regions; in finance, funds rely on predicting future stock return rates to continually adjust their portfolios and maximize expected returns; and in wildlife conservation, patrol decisions to prevent poaching are based on predictions of poachers' behavior. To address underlying uncertainty issues, this project develops new methods for tailoring predictive models to capture the underlying uncertainty factors and integrate them into decision-making problems. This research will develop a general decision-focused learning framework, which will be applied in public health, environmental sustainability, and urban planning. Furthermore, this research will foster a diverse group of doctoral and undergraduate students at Georgia Tech who collaborate across disciplines, and create a graduate-level course focused on machine learning for decision-making at Georgia Tech. Part of the uncertainty issues incurrent decision-focused learning methods is based on data scarcity issues, due to challenges in task-specific training protocols, model misspecification, and data distribution shifts during deployment. This project addresses the pressing need for data-efficient decision-focused learning that can effectively solve stochastic decision-making problems (i.e., problems with high levels of uncertainty) with limited data, while addressing the key challenges of model misspecification and quickly adapting to changing distributions in the deployment phase. It will create a pre-training and fine-tuning learning paradigm that modularizes the key components in decision-focused learning for stochastic optimization. The specific objectives include: (1) pretraining key components to obtain general task-agnostic base modules that can be reused in different tasks; (2) fine-tuning the pretrained modules for specific decision-making tasks rapidly using task-specific data; and (3) developing active data collection techniques using model uncertainty to improve model performance and adapt the model to changing distributions. By exploiting broader data across multiple tasks during pre-training and requiring only a small amount of task-specific data for fine-tuning, this framework creates flexible predictive models that reduce model misspecification, improve data efficiency, and mitigate distribution shifts during deployment to produce better decisions. These research thrusts compose a general decision-focused learning framework. 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 2024 · 2024-09

Recent observations by the IceCube Neutrino Observatory have provided evidence for the first two candidate sources, powered by supermassive black holes in other galaxies, as well as the emission of a high-energy flux within our own Milky Way galaxy. The Pacific Ocean Neutrino Experiment (P-ONE), a leading next-generation neutrino observatory located deep in the Pacific Ocean off the coast of Washington State, is under development to advance this nascent field of neutrino astrophysics. This award will support a team of researchers at the Georgia Institute of Technology, Drexel University, Michigan State University, and the University of Chicago to develop and construct a start-of-the-art trigger system for the P-ONE observatory that will collect and process all the P-ONE data. This award will provide unique skill development in advanced scientific electronics and data acquisition for early career scientists, including students and postdoctoral researchers. The new observatory is a planned integral part of the global multi-messenger astrophysical observations program that includes many NSF-funded instruments, and it is developed in partnership with international collaborators. More broadly, the data facilitated by this trigger system will provide continuous monitoring of environmental conditions and biological processes in an unprecedentedly large volume of the deep ocean, providing novel input to the oceanography and marine biology communities. A central challenge in neutrino astrophysics is identifying discrete sources and associating them with astronomical objects. This is currently limited by the angular resolution of existing neutrino telescopes, such as IceCube. By leveraging the optical properties of deep-sea water and modern electronics, P-ONE can improve angular resolution by a factor of at least 4, resulting in better sensitivity by the same factor. The development of the trigger system under this award will provide the crucial link between raw detector data and analysis, delivering all scientific results from the detector, that is essential for realizing the scientific goals of the P-ONE. The trigger system that will be developed under this award is central to the future full cubic-km-scale P-ONE observatory, ensuring a central role for the US and will assure US leadership in the emerging field of high-energy neutrino astrophysics. 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.