Song Han, Ph.D. - US grants

Increasing use of energy efficient home appliances contributes to the reduction of energy usage by residential consumers. The combination of home appliances and energy management systems has been proposed as an energy efficient solution for smart homes. However, the need of reactive power not only reduces overall energy efficiency due to losses incurred in transmission and distribution power lines, but also poor regulation of reactive power degrades the lifespan of consumer appliances; both quantifiable in billions of dollars in losses per year. In this project, the available reactive power capacities in various home appliances will be evaluated. The reactive power demand will be met by the reactive power capacity of these appliances. Typically, home appliances such as laundry machines and air conditioners have a front-end power factor correction (PFC) converter to satisfy power factor conditions for Energy Star certification. These PFC circuits are usually low-cost unidirectional boost converters. Even though the power capacity in this PFC converter is small, multiple home appliances can cooperate to meet the reactive power demand in distribution level power networks. A wireless home power management system will also be developed to 1) manage the reactive power compensation provided by each appliance, and 2) cooperate with smart home energy management networks or advanced metering interface systems for peak demand response purposes. As a result, the power losses due to the transmission lines and distribution lines, and voltage fluctuations due to variations in reactive power will be significantly reduced. The success of the proposed framework will result in significant enhancement of power quality in residential applications at little to no extra cost to the consumer. Furthermore, the life-cycle of the home appliances will be improved; and lastly, CO2 emissions will be reduced.

The objective of this research is to construct, emulate, and investigate the effectiveness of a wireless power quality management system that utilizes available power capacities from home appliances, small solar power generators, and electric vehicles which can all be found in residential and smart home applications. The proposed method coordinates active and reactive power support and harmonic control capabilities of both unidirectional converters and bidirectional converters for power quality mitigation. A low-power real-time wireless network will be employed as the communication infrastructure to exchange power quality information between a power quality manager and the converters. Hybrid control for active power, reactive power, and harmonic distortion compensations in these aggregated converters will be the key enabler for this technology. This project will conduct a technology survey, wireless network and supervisory controller design, an experimental verification through proof-of-concept design, and an implementation and feasibility test of reactive power support from home appliances. The proposed research is transformative because it provides a fundamental and highly scalable framework for monitoring, mitigating, and enhancing power quality in residential applications ranging from a single building power system to community-wide and city-wide power systems especially in microgrids employing renewable energy sources, power quality monitoring systems, and smart appliances.

Discovering latent subgroups in a sample is an important problem in many scientific disciplines. Social scientists identify subgroups within a population based on behavioral patterns to examine differential effects of social status. Engineers recognize malfunctions of a manufacturing system based on performance measures to detect design defects. Physicians define subtypes of a disorder on the basis of clinical symptoms to identify associated genetic risk factors. This kind of problem involves two sets of variables: a set of descriptors describing the issue (e.g., behavioral patterns, or symptoms) and a set of moderators or predictors (e.g., social status, or genetic factors). The ability to accurately predict the latent classes (e.g., disease subtypes) from predictors (e.g., genetic risk) in the absence of observed descriptors (e.g., before symptoms are developed) will advance many of these disciplines. This project aims to develop an effective and efficient platform of machine learning algorithms to solve this problem. The team will effectively integrate research and teaching to engage students into the proposed study. Validated methods and software will be broadly disseminated through the project web repository and scientific presentations.

This project addresses the latent class discovery and prediction problem by deriving novel and efficient approaches, including multi-view co-clustering, multi-view subspace clustering, multi-objective optimization of co-training, and multi-modal deep learning methods. Parallel and distributed algorithms will be developed to implement and scale up these methods. A streamlined analytics platform will be constructed to maximize the utility of the proposed approaches in real-world applications. The proposed solutions will be evaluated in the analysis of large-scale sensory and behavioral data. By collaborating with domain experts, the project will (1) identify risk factors for problematic human behaviors such as binge drinking; and (2) locate the sensory features most discriminative of gait abnormalities due to neurological disorders such as Parkinson's disease or stroke.

Urban areas, where more than 80% of US population lives and 80% of energy is consumed, are developing into smart, connected communities. At the heart of city infrastructures is the urban power distribution network, which supports various systems including government, safety, water, food, transportation, communication, and other functions vital to the lives and work of citizens. Current urban distribution networks were not designed for smart cities, and cannot sustain the ever-increasing demands from urban growth in the face of substantial increases in renewable generation and extreme weather-induced blackouts. Lack of a scalable and high-speed communication and computing infrastructure is a key bottleneck. This project will architect a novel Software-Defined Distribution Network (SD2N), a gigabit networking and computing platform to enable a sustainable and resilient electric power Internet for smart cities. SD2N will manage a vast number of smart grid devices, allow self-adaption, self-management and self-healing without costly hardware upgrades, and provide a sustainable, scalable and replicable smart city backbone infrastructure. The new architecture will illustrate how software-defined networking and distributed real-time computing can provide urban infrastructures with resilient, sustainable, human-centered, highly efficient and affordable service platforms for smart cities. The concepts and platform will have the potential to be applied across industry sectors, with significant benefits to municipalities, utilities, developers, and their stakeholders. The proposed SD2N platform will be demonstrated at US Ignite Summits, and the research results will be transferred to key stakeholders, communities and cities in collaboration with Eversource Energy and the Connecticut Center for Advanced Technology.

The innovation of the project lies in integrating Internet of Things technologies, software-defined networking and real-time computing to establish a scalable SD2N architecture. It will combine a hybrid software-defined networking infrastructure and a distributed real-time computing framework with advanced optimization to enable self-configuration, scalable monitoring, real-time data streaming, processing, storage and feedback while tackling the stringent data availability and multi-latency requirements in managing urban distribution networks. The proposed SD2N architecture will enable coordinated economic dispatch of microgrids to significantly reduce their carbon footprints and total operation costs while supporting smooth, grid-friendly renewable energy penetration. It will also enable shared electricity services among connected communities, leading to resilient energy service for smart cities. An SD2N prototype, to be established in the University of Connecticut's innovative hardware-in-the-loop real-time test bed, will offer valuable resources for research communities, as well as the energy and IT industries.

In the next generation of big science experiments, the demands for computing resources are expected to outstrip the capabilities of existing computing infrastructure. In light of this, a radical rethinking of the cyberinfrastructure is needed to contend with these developments. With the onset of deep learning, parallelized processing architectures have emerged as a solution. Combined with deep learning algorithms, parallelized processing architectures, in particular, Field Programmable Gate Arrays (FPGAs) have been shown to give large speedups in computing when compared with conventional CPUs. This project aims to bring machine learning based accelerated computing with FPGAs into the scientific community by targeting two big-data physics experiments: the Large Hadron Collider (LHC) and the Laser Interferometer Gravitational-wave Observatory (LIGO). This project will push the frontiers of deep learning at scale, demonstrating the versatility and scalability of these methods to accelerate and enable new physics in the big data era. This project serves the national interest, as stated by NSF's mission, by promoting the progress of science. The PIs and their collaborators will build upon their recent work to design and exploit state-of-the-art neural network models for real-time data analytics, reducing overall computing latency. This new computing paradigm aims to significantly increase the processing capability at the LHC and LIGO, leading to an increased scientific output of these devices and, potentially, foundational discoveries. The students to be mentored and trained in this research will interact closely with industry partners, creating new career opportunities, and strengthening synergies between academia and industry. In addition to sharing algorithms with the community through open source repositories, the team will continue to educate the community regarding credit and citation of scientific software.

In this project, the PIs will build upon their recent work developing high quality deep learning algorithms for real-time data analytics of time-series and image datasets using Field Programmable Gate Arrays (FPGAs) to accelerate low-latency inference of machine learning algorithms. The team will develop machine learning based acceleration tools focusing on FPGAs to be used within LIGO and the LHC experiments. The team's immediate goal is to take benchmark examples of LHC high level trigger processing and LIGO gravitational wave processing and construct demonstrators in each scenario. For this benchmark, they aim to design and implement an FPGA based accelerator that can perform low latency gravitational wave identification and LHC event reconstruction. Additionally, the PIs aim to add the capability of graph based neural network accelerators for FPGAs. The open source tools to be developed as part of these activities will be readily shared with LIGO, LHC, and LSST. The project will create an advisory group, including members of large and small projects, members of the neutrino physics, multi-messenger astronomy community, industry partners, computer scientists, and computational biologists. This project aims to bring together representatives of the different communities that will benefit from and can contribute to this work. The PIs will organize deep learning workshops and boot camps to train students and researchers on how to use and contribute to the framework, creating a wide network of contributors and developers across key science missions. This project is part of the National Science Foundation's Harnessing the Data Revolution (HDR) Big Idea activity.

This project is part of the National Science Foundation's Harnessing the Data Revolution Big Idea activity. The effort is jointly funded by the Office of Advanced Cyberinfrastructure.

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.

Industrial Internet of Things (IIoT) systems are used in a wide range of mission- and safety-critical applications, thus imposing stringent requirements on the security of the underlying communication infrastructure. An IIoT network consists of multiple communication parties and follows a two-way communication model, including delivering sensing data on the uplink and transmitting control messages on the downlink. Tampered sensing data or control messages by outside attackers will result in wrong decisions, potentially causing significant harm. The recent trend in industrial automation to connect interdependent industrial plants together to provide decentralized, verifiable and immutable services further exacerbates the problem. This project aims to design 1) efficient signature schemes to support verifiable authenticity, integrity, and uniformity for intra-plant two-way communications, and 2) hierarchical and scalable blockchain protocols to support inter-plant immutable services. The close collaboration of the research teams will lead to a publicly available IIoT-enabled advanced manufacturing testbed, effective dissemination of research results among practitioners, and initiation of technology transfer.

To address existing limitations, the proposed secure communication framework aims to (i) ensure authenticity, integrity, and uniformity of sensing data in IIoT networks by designing novel signature schemes that are fast and efficient for both the signer and the verifier; (ii) enable public-key cryptography (PKC)-based fast control message authentication by extending the control border of IIoT networks to the cloud/Internet and solving the new security challenges; and (iii) provide inter-plant immutable services by developing a hierarchical blockchain structure and scalable lightweight consensus protocols. The proposed solutions will be implemented and deployed on a unique IIoT-enabled advanced manufacturing system testbed for thorough design validation and performance evaluation. Successful design, implementation and demonstration of the proposed security solutions should advance the adoption of IIoT network infrastructure, accelerate the transformation of legacy security architectures to PKC-based security architectures and lift the security protection of the industrial communication infrastructure to the next level.

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.

The broader impact/commercial potential of this Partnerships for Innovation - Technology Translation (PFI-TT) project is to enable the process industries to fully embrace the concept of Industry 4.0 and revamp their manufacturing plants as smart factories, where intelligent sensing and field devices are pervasively deployed and wirelessly connected in the field to provide real-time high-speed process monitoring, diagnosis and control. This paves the way for a better understanding of the manufacturing process, thereby enabling efficient and sustainable production. The commercial availability of low-cost, real-time and high-speed wireless technology will enable a wide range of new applications benefiting society in many ways. Just as ordinary WiFi technology has transformed the landscape of wireless communication for business use, our new real-time version of WiFi will enable industrial applications such as advanced process control and manufacturing, assistive tele-robotics, high-bandwidth health monitoring, autonomous vehicle safety assurance, and others. This PFI-TT project will also enable a comprehensive training program for future leaders in innovation and entrepreneurship to develop their ability to conduct innovative research projects as well as effective communication and entrepreneurial skills, and thus have significant educational and societal impacts.

The proposed project will address the drawbacks of existing wireless communication networks, which are not equipped with providing either deterministic or high-speed real-time communication for large-scale industrial sensing and control systems. The proposed RT-WiFi protocol will 1) support up to 6kHz sampling rates that is sufficient for most industrial wireless applications, which require highly predictable performance; 2) offer the capability for extending to multi-cluster settings to cover large factories; 3) provide extremely high reliability through dynamic and distributed packet scheduling for autonomous adaptation to the noisy factory environment; and 4) tolerate any single point hardware/software failures with robust gateway design. It is envisioned that upon the successful completion of this project, a full-blown RT-WiFi communication system can be developed to support a large number of reliable and dynamic real-time flows in large-scale high-speed industrial wireless control networks. These unique features will make RT-WiFi an ideal communication infrastructure to support not only advanced process control systems, but also other fast growing industry segments in robotics, autonomous vehicles and high-speed mechanical processes.

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.

An Industrial Internet of Things (IIoT) paradigm aims at creating unified sensing, computing, and control framework to interconnect all the industrial assets with information systems and business processes and to streamline the manufacturing process and lead to optimal industrial operations. Because IIoT applications - including autonomous driving and smart highway, manufacturing automation with robots, etc. - are distinguished from commercial IoT by stringent performance guarantees and certifiable robustness, research is needed to provide a holistic resource management framework that enables effective sensing and control operations in the presence of intermittent data sources and unpredictable system disturbances. This project aims to lay the foundation for such a framework by formulating and investigating three fundamental questions: 1) How to achieve real-time data retrieval with intermittent data sources and large-scale high-speed wireless control with guaranteed performance? 2) How to perform dynamic packet scheduling to compensate for unexpected system disturbances? 3) How to perform composite resource management to jointly consider network and computing resources for resource scheduling among multiple IIoT applications? By addressing these questions, the proposed dynamic and composite resource management framework has the potential to vastly advance the adoption of IIoT technologies, accelerate the transformation of legacy communication infrastructure to advanced wireless infrastructure and boost the nation's economic growth and competitiveness.

To fundamentally transform the design principles of resource management in large-scale IIoT systems, this project will (i) design novel algorithms for real-time data management in IIoT systems with intermittent data sources; (ii) develop new scheduling techniques for control performance optimization in multi-cluster wireless networks; (iii) design a fully distributed packet scheduling framework to handle unexpected system disturbances in complex industrial environments; and (iv) explore new models and scheduling methods to develop a composite resource management framework for handling heterogeneous resource scheduling, partitioning and reconfiguration for time-critical end-to-end services in large-scale IIoT systems. These innovations will be validated using high-fidelity IIoT simulation tools and deployed on university-industry co-established IIoT testbeds for thorough performance evaluation. This proposed resource management framework will provide researchers and industrial partners holistic solutions to achieve provable performance in large-scale IIoT systems and support a wide range of industrial applications. The research outcomes will be integrated into an innovative professional education program at the University of Connecticut to educate current and next-generation researchers and professionals in a creative way to understand, appreciate and contribute to the fast-growing and rapidly evolving IIoT technologies.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Personalized, precision healthcare (PPH) utilizing edge sensing-computing can collect, analyze and interpret continuous, multi-modality data, both physical and physiologic, producing information, knowledge and insight needed for real-time disease onset and progression monitoring at both the individual and population levels. This planning proposal will (i) identify the challenges and investigate the principles and potential solutions for the edge sensing-computing paradigm; (ii) engage diverse academic, community and government stakeholders to collectively define the functional and performance requirements for PPH; and (iii) create and validate preliminary approaches and devise a concrete, detailed plan for scaling PPH to national levels. It is well aligned with NSF?s mission to ?advance the national health, prosperity and welfare.? This project can generate enormous social and economic benefits for communities, healthcare systems, and other stakeholders. If successful, the project will enable the monitoring of epidemics (e.g. disease outbreaks/spread, early detection/preemptive intervention of acute/infectious diseases) and the management of chronic physical and psychological conditions. The PIs will 1) disseminate publications, data and systems in academic, industry and community venues; 2) integrate CISE student education (including female and under-represented minorities) at different levels; 3) mentor high-school students on joint health-technology research; 4) cultivate a technology-literate healthcare workforce; and 5) pilot the technologies for immediate benefits to nearby communities while studying how to scale to other rural, suburban, and city settings.

This project will explore, design, and evaluate potential solutions for enhancing the scalability of edge sensing-computing-based PPH in four dimensions of different types of sensing data, analytic algorithms, diseases, health conditions, and population sizes. The PIs will identify challenges and validate approaches guided by three principles: privacy as a first-class citizen, design for faults and exploitation of scale. The team will: 1) define new abstractions and quantifiable metrics for end-to-end security and privacy guarantees across hardware, software and application stack; 2) investigate systems for multi-temporal resolution processing of heterogeneous healthcare data, incorporating composition of components from possibly untrusted third parties and accommodate noises, disturbances or even adversary-controlled data; 3) explore novel AI/machine-learning algorithms suitable for PPH learning and inference, including AutoML for neural-network architecture search, model compression and federated learning at extreme scale while meeting security, privacy and robustness constraints; and 4) develop heterogeneous hardware accelerators and general design methodologies and tools for neural-hardware architecture co-design, efficient acceleration for time-series, point-cloud and language/sound understanding, and on-device edge training.

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.

The growing capabilities of sensing, computing and communication devices are leading to an explosion of Internet of Things (IoT) infrastructures. In the meantime, advances in technologies such as autonomous systems and artificial intelligence promise enormous economic and societal benefits. Naturally, it is desirable to deploy these technologies in IoT infrastructures. However, such deployments present daunting changes for increasingly scaled-up IoT infrastructures in mission-critical applications such as medical, energy, transportation, and industrial-automation systems. These challenges stem from several major aspects in terms of scalability. First, the number of edge devices can be enormous, often in the order of billions, which makes centralized management infeasible. Second, there are multiple layers of heterogeneity. An IoT system can consist of heterogeneous computing subsystems; each subsystem can have heterogeneous computing devices; and each single device can be composed of different kinds of computing components. Third, mission-critical applications have stringent requirements in correctness, resilience, timeliness, security and safety. It is difficult for a large-scale IoT system to satisfy these requirements due to increasing opportunities for adversarial activity.

To tackle these challenges, this project aims to develop a cross-layer and full hardware/software stack solution, referred to as the S3-IoT framework, for the design and deployment of scalable, secure, and smart mission-critical IoT systems. The S3-IoT framework will span three different computation layers, including data centers, gateways/aggregators, and edge devices, and cover four research foci, i.e., resource management, security and privacy, computer architecture/systems, and algorithms. In this planning project, an initial version of the S3-IoT framework will be developed. The S3-IoT framework will (i) leverage a layered structure - data centers, gateways/aggregators, and edge devices to accommodate the huge number of edge devices; (ii) develop cross-layer techniques to deal with the heterogeneity among these layers; and (iii) propose hardware and software co-design approaches that embrace the heterogeneity among computing components to improve the performance of all components within an individual layer. The S3-IoT framework will be evaluated by developing simulators with the layered structure as well as a small scale, comprehensive experimental testbed. The success of this planning project will lead to a convincing path to effective deployment of mission-critical IoT systems and infrastructures, particularly in terms of improving resilience to environmental uncertainties, system internal errors and faults, and malicious attacks.

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.

The novel coronavirus, COVID-19 is a pandemic infecting people in the United States and around the world. It is of utmost importance to prevent the fast spread of the virus. This project will use artificial intelligence (AI) methods to slow down the infection by encouraging proper wear of Personal Protective Equipment (PPE) by hospital staff and by supporting social distancing. The planned method will help monitor dangerous activities pointed out by Center for Disease Control (CDC), such as hand-to-face contact, touching inside or crossing arms when taking off the gown and masks and social distancing. It will advance the national health, protect the healthcare workers and help the whole nation combat the pandemic.

Video understanding and activity recognition have made great progress in recent years. This project will apply artificial intelligence on a mobile platform for efficient activity recognition techniques to guide people's activities in healthcare settings, including patients', health care workers' and community residents'. To protect privacy in transmission to the cloud, the project applies the team's work on model compression techniques and neural architecture search to make the AI more compact and efficient so that it can be deployed on edge devices. As a result, videos can be locally processed; only key information or the detection result is sent over the cloud, preserving people's privacy. Finally, the project will efficiently deploy such algorithms on mobile devices and make it freely available in the hospitals.

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.

This project investigates a completely new cross-disciplinary concept of “Computational Screening and Surveillance (CSS)” that utilizes edge learning to detect early indicators of diseases, and monitor health changes in both individuals and populations. CSS analyzes and interprets continuous and heterogeneous physical and physiologic sensing-data streams of human subjects to produce real-time information, knowledge, and insights about their health status. The project’s novelty is a data-driven paradigm that revolutionizes the understanding, prediction, intervention, treatment, and management of acute/infectious, chronic physical and psychological diseases. The project’s impacts are enormous social and economic benefits to individuals, organizations, and the healthcare system: early detection, preemptive intervention and management can lead to greatly improved quality of care, and huge savings for multiple diseases each costing hundreds of billions of dollars every year.

The investigators design, develop and evaluate principles and solutions for CSS enabled by extreme-scale edge learning spanning four dimensions: data modalities, health conditions and data patterns, Artificial Intelligence/Machine Learning (AI/ML) algorithms and models, and individuals/populations. The design is guided by four principles: exploit scale and heterogeneity, design for uncertainty, privacy as a first-class citizen, and faults and attacks as a norm. The investigators will 1) design AI/ML algorithms for learning data patterns and correlations for diverse health conditions in both individuals and populations at extreme scales; 2) quantify theoretical bounds on the tradeoffs between security, privacy protection, and learning accuracy in order to protect against various attacks on data and models at both the edge and cloud; 3) develop programming abstractions for automated exploration of competing AI/ML methods under uncertainty, and system mechanisms to protect stream processing integrity against sensitive data disclosure and faulty/malicious analytics; and 4) devise neural architectures and accelerators for computation efficiency at the constrained edge, data efficiency using limited training sets, and human efficiency utilizing AutoML.

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.

The recent breakthrough of on-device machine learning with specialized artificial-intelligence hardware brings machine intelligence closer to individual devices. To leverage the power of the crowd, collaborative machine learning makes it possible to build up machine-learning models based on datasets that are distributed across multiple devices while preventing data leakage. However, most existing efforts are focused on homogeneous devices; given the widespread yet heterogeneous participants in practice, it is urgently important but challenging to manage immense heterogeneity. The research team develops heterogeneous architectures for collaborative machine learning to achieve three objectives under heterogeneity: efficiency, adaptivity, and privacy. The proposed heterogeneous architecture for collaborative machine learning is bringing tangible benefits for a wide range of disciplines that employ artificial intelligence technologies, such as healthcare, precision medicine, cyber physical systems, and education. The research findings of this project are intended to be integrated with the existing courses and K-12 programs. Furthermore, the research team is actively engaged in activities that encourage students from underrepresented groups to participate in computer science and engineering research.

This project provides the theoretical underpinning and empirical evidence for an efficient, adaptive and privacy-preserving design under heterogeneity, which fills a critical void - the existing collaborative machine-learning approach fails to manage the immense heterogeneity in practice. This project centers on three aspects: (1) design of specialized neural architectures for heterogeneous hardware platforms to cope with the limited efficiency of collaborative training due to heterogeneity; (2) design of an efficient and adaptive knowledge-transfer framework to bridge heterogeneous participants based on their underlying proximity benefits; (3) privacy strategies for heterogeneous collaboration by identifying new vulnerabilities and developing privacy-preserving mechanisms. A general-purpose testbed is built to rigorously validate the proposed research and expand the impact of this project. It is expected that this project opens a new research paradigm to unleash the utmost potential of heterogeneous and collaborative machine intelligence.

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.

The data revolution is dramatically accelerating the acquisition rate of new information, creating a vast amount of data. Artificial intelligence (AI) has emerged as a solution for rapid processing of complex datasets. New hardware such as graphics processing units (GPUs) and field-programmable gate arrays (FPGAs) allow AI algorithms to be greatly accelerated. To take full advantage of fast AI, the Institute of Accelerated AI Algorithms for Data-Driven Discovery (A3D3) targets fundamental problems in three fields of science: high energy physics, multi-messenger astrophysics, and systems neuroscience. A3D3 works closely within these domains to develop customized AI solutions to process large datasets in real-time, significantly enhancing their discovery potential. The ultimate goal of A3D3 is to construct the institutional knowledge essential for real-time applications of AI in any scientific field. Through dedicated outreach efforts, A3D3 will empower scientists with new tools to deal with the data deluge. Students mentored through A3D3 research will interact closely with industry partners, creating new career opportunities and strengthening synergies between academia and industry.

The approach of A3D3 is to tightly couple AI algorithm innovations, heterogeneous computing platforms, and science-driven application development informed through close collaboration with domain scientists within physics, astronomy, and neuroscience. The common theme across domains is the development of AI strategies accelerated by emerging processor technology, employing hardware-AI co-design as a transformative solution to a wide range of scientific challenges. Hardware architectures such as GPUs and FPGAs have emerged as promising technologies to address many of the challenges in data-intensive science because they provide highly-performant, parallelizable, and configurable data processing pipeline capabilities. When combined with AI algorithms, these architectures significantly accelerate scientific workflows compared to CPU-only computing platforms. Building on the existing Fast Machine Learning community, A3D3 cultivates an ecosystem where scientists across domains collaborate to meet critical challenges, forming a central hub of excellence for innovation in accelerated AI for science. The work is extended to the public at large through a diverse set of educational training programs and by mentoring next-generation scientists.

This project is part of the National Science Foundation's Big Idea activities in Harnessing the Data Revolution (HDR) and Windows on the Universe - The Era of Multi-Messenger Astrophysics (WoU-MMA). This award by the Office of Advanced Cyberinfrastructure is jointly supported by the Divisions of Astronomical Sciences and of Physics within the NSF Directorate for Mathematical and Physical Sciences.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.