In this proposal, we aim to design, build, and integrate a novel student modeling engine to provide adaptive scaffolding to programming students. We bring in our expertise in student modeling coupled with LLM's capability in deep analysis of programming snippets to address challenges associated with the temporal ill-defined nature of programming education. This entails effective representation of students' programming processes and capturing their evolving competency in foundational programming knowledge and skills.

The ever-increasing surge in interest in computer science (CS) education, coupled with the unprecedented changes brought about by the emergence of generative programming tools, underscores the imperative for educational advancements in this field. These advancements may include understanding the most effective approaches to support students' learning integrated within state-of-the-art adaptive technologies that can provide CS students with effective, scalable, and individualized educational support. Our proposal centers on the creation of a personalized multi-level programming practice environment. This will be achieved by leveraging models of students' learning, a vast corpus of worked examples and practice problems, and the latest advancements in machine learning and generative AI. Through this proposal, we aim to conduct extensive studies to discern the most effective types of programming exercises for enhancing students' learning, taking into account their practice history and learning trajectories. The insights gained from these studies will be seamlessly integrated into our personalized CS education learning environment. This integration will result in tailored exercises that match each student's competency level, accompanied by individualized support to guide them through the exercises. To validate the effectiveness of our approach, we plan to evaluate it in several introductory programming classrooms across a variety of institutions. These institutions include large state universities and small HBCU colleges, ensuring that our solution addresses the diverse needs of students across different educational settings.

This project directly contributes to the main goals of the NSF DRK-12 program through the design, implementation, and evaluation of a project-based integrated science and AI curriculum and technology. The project has the overarching goal of preparing a diverse, computationally-competent next-generation STEM workforce through devising an effective, engaging, and inclusive learning environment. In particular, we will create a learning environment for computational modeling that features custom code blocks that facilitate the implementation of AI algorithms in science-related contexts while obfuscating unnecessary programming complications. Our curricular modules will integrate important AI concepts with the high school STEM curriculum following NGSS standards.

With the rapidly growing recognition of the role that computer science is playing in every aspect of society, enrollments in introductory computer science courses are increasing at an unprecedented pace. As a result of this phenomenal growth, departments of computer science are seeing extraordinary demand for introductory computer science courses. The accelerating growth in enrollments poses significant challenges for introductory programming instructors, who must teach increasingly larger classes while providing effective, engaging learning experiences for students. The overarching objective of this project is to develop an introductory programming teaching support environment, INSIGHT, that will enable instructors to readily understand their students� progress through introductory computer science coding activities. INSIGHT will fundamentally change classroom dynamics by supporting both students and instructors.

Exploring Computation Integrated into Technology and Engineering II (ExCITE II) is a four-year High School (HS) Strand Large CSforAll:RPP that builds upon the work of the successful ExCITE I medium scale HS strand CSforAll:RPP project (NSF 1923552). ExCITE II will use the Beauty and Joy of Computing (BJC) curriculum (NSF 1138596 and 1441075), which was enhanced during ExCITE I (to produce the Advanced Placement (AP) Computer Science Principles (CSP) by Design version) by the addition of hands-on real world problem-solving activities with external computer control/robotic devices, to enliven students’ interest in the big ideas of computing (see p. 7). ExCITE I methodologies, research results, and tested AP CSP by Design curriculum materials will be used to anchor a full spectrum of professional development (PD) workshops for 608 Technology and Engineering (T&E) educators who serve diverse students in 12 states―as a prototypical model for enabling T&E teachers nationwide to provide pedagogically sound AP CSP instruction.

This project will support implementation and study of the Beauty and Joy of Computing (BJC) curriculum. We aim to increase implementation of BJC in New England states and beyond particularly in high-need districts. We will study the effects of BJC implementation on the participation of girls, Black, Latinx and low-income students.

The STARS Computing Corps Alliance addresses the challenge of increasing the number and representation of Black, Hispanic, and women students who graduate with computing degrees and who remain in the field of computing after graduation. The primary goal of the alliance focuses on retaining students in computing degree programs in higher education: STARS aims to increase persistence and degree attainment rates in computing degree programs for Black, Hispanic, and women students to be at parity with their representation in the broader population. Building on the prior success of the STARS Computing Corps Alliance for Broadening Participation in Computing, the goal of the this project is to significantly extend the STARS alliance to expand upon those impacts, by 1) including new partners that expand the reach of STARS and that emphasize participation of Black and Hispanic students and faculty, particularly from emerging Hispanic Serving Institutions and community colleges, 2) creating new program elements that test new and propagate evidence-based BPC practices within computing departments, and 3) leveraging partnerships to support K12 outreach that applies best practices for broadening participation in computing and enhances learning opportunities for K12 students and teachers around AI concepts, and 4) developing STARS Alumni groups employed in industry positions to promote transition to and retention within the tech workforce.

To address challenges teaching CS in middle school classrooms, we need innovations in teacher professional development for computational thinking and advances in real-time teacher support tools that empower teachers to make data-driven instructional decisions to meet the needs of all learners. Leveraging recent advances in artificial intelligence (AI) and building directly on recent work by the project team, this project will address the following goals: (1) provide CT learning experiences to teachers by advancing and implementing professional development (PD) for computationally enriched science. We will investigate the ways in which these learning experiences improve instructional practices by fostering teachers’ self-efficacy, content knowledge, and pedagogical knowledge for leading CT in science; and (2) develop an AI-augmented toolset that supports teachers in making data-driven instructional decisions during computationally enriched science activities. This toolset, TRACES, will be co-designed with teachers to provide real-time information that supports teachers in noticing and responding to student work.

Unmanned Aircraft Systems (UASs), or drones, have tremendous scientific, military, and civilian potential for data collection, monitoring, and interacting with the environment but don't have control and planning algorithms capable of adjusting performance as computing resources are continually reallocated, such as when transitioning from waypoint navigation to environmental sample collection. A computing framework to make use of freed resources will be developed allowing autonomous agents to focus attention where it is needed, for example, away from navigation and to perception. Together, these will provide a blueprint for making use of similar algorithms with adjustable performance (e.g., anytime algorithms) which can be adapted to other robotics platforms, as well as water, space, or ground vehicles. These technology innovations will improve the ability of agents to learn more, perceive more accurately, collect better data, and respond more appropriately to changing environments and mission objectives. Specific to UAS, this project will help maintain U.S. air superiority goals through agile planning, targeted and persistent Intelligence, Surveillance, and Reconnaissance (ISR), and flexibility and adaptability. The project goals are coupled with outreach and educational activities focused on increasing the understanding of rural populations of the value of investing in scientific and technological research. The educational efforts, targeted at K-12, undergraduate, graduate, and adult engagement are designed to dramatically increase the CPS educational pipeline in the Midwest.

The proposed REU site will provide undergraduates with a comprehensive research experience in the context of unmanned systems. We will provide systematic and broad instruction in research methodology to help students become independent researchers, focused research activity through ongoing research projects on a variety of unmanned systems.Targeted student participants include computer science, computer engineering, mechanical engineering, electrical engineering, and other related majors. Particular focus will be given to students from historically underrepresented groups and from institutions in more rural communities that lack research opportunities.

For many forms of interactive environments, the system's behaviors can be viewed as a sequential decision process wherein, at each discrete step, the system is responsible for selecting the next action to take from a set of alternatives. The objective of this CAREER proposal is to learn robust interaction strategies that will lead to desirable outcomes in complex interactive environments. The central idea of this project is that strategies should not only be effective in complex interactive environments but they should also be efficient, focusing solely on the key features of the domain and the crucial decision points. These are the features and decisions that are not only associated with desirable outcomes, but without which the desirable outcomes are unlikely to occur.

This project will develop generalizable data-driven tools that addresses the conceptually and practically complex activity of constructing adaptive support for individualized learning in STEM domains.

Despite the fact that there are approximately 370 food banks in the US, rescuing and distributing billions of pounds of food annually, there is still a consistent and persistent number of people struggling daily to obtain access to safe and nutritious food. The challenges non-profit hunger relief organizations (NPHRO) face have only been exacerbated by the COVID-19 pandemic. The food insecure clients have experienced challenges such as waiting in long lines to obtain food; potentially being turned away as food banks run out of food. In contrast, hunger relief organizations, particularly during extreme events like a pandemic have to grapple with food sourcing and access issues such as (i) obtaining in-kind donations given a more constrained supply chain environment; (ii) procuring food to mitigate in-kind donation shortages given rising food prices, stock-outs, and long lead-times for some of the items; (iii) doing more with fewer resources given shortage of volunteers and truck drivers due to social distancing and quarantining; and (iv) moving to contactless delivery operations. In essence, these organizations operate a supply chain that has to be adaptive and responsive in an environment characterized by multiple decisions, multiple supply chain actors, and complex information about supply, demand, and the physical processes connecting supply acquisition to distribution. However, much of the intelligence required to perform what-if analyses and forward looking assessments of supply, equitable distribution, agency recruitment, and capacity expansion, and resource allocation are relegated to spreadsheet assessments combined with information extracted from other sources and user domain knowledge. This PFI-RP proposal seeks to support the higher-level scenario analyses and decision-making that NPHROs need through the development and potential commercialization of a smart food distribution system for resource allocation under extreme events. We specifically seek to build upon a prior grant where a flexible, equitable, effective, and efficient food distribution prototype was developed to support normal operations of a food bank organization; namely predict in-kind donations over time and use these predictions to facilitate distribution within the service area at the county level that considers the tradeoff between equity, effectiveness and efficiency in an interactive visualization.

Knowledge graphs have emerged in many domains of science and technology as a powerful means of integrating, structuring, and mining information to extract new knowledge. Recognizing the importance of this paradigm, the Proto-OKN project will create FRINK: Fabric Integrating Networked Knowledge. FRINK will create capabilities that allow for the uniform deployment, integration, and harmonization of knowledge graphs created under the Proto-OKN program into a unified Open Knowledge Network for query and analysis. FRINK will be organized around three objectives, detailing three types of fabric that will bind together the initially disparate graphs developed by Theme 1 teams.

Autonomous Driving Systems (ADS) are software systems designed to reduce the need or even replace humans in the task of driving a vehicle. They have been attracting tremendous societal interest given its potential to increase road safety, reduce traffic congestion, reduce commuting time, etc. Unfortunately, these systems are not foolproof. Physical testing of ADS-driven vehicles does not scale and can be unsafe. For those reasons, simulation-based testing (SBT) became very popular. Despite the recent advances in SBT techniques for autonomous driving, important challenges remain to be addressed. First, finding problematic inputs with simulation is time and space costly. Second, techniques often report duplicate infractions that do not contribute to information gain. Third, techniques make strong assumptions about the simulation environment. This project proposes novel approaches to mitigate these fundamental challenges, advancing the state of the art in Simulation-based Testing of Autonomous Driving Systems.

Runtime Verification (RV) monitors programs against formal specifications and reports violations when executions do not satisfy those specifications. RV can find bugs that tests miss, but it is not yet widely adopted. We address two hindrances to broad RV adoption: (1) writing specifications requires learning domain specific languages, and (2) many programming languages have no mature RV tool. We propose infrastructure for writing specifications in popular user-friendly formats, and for reusing existing RV tooling for Java to monitor programs written in other languages. We will evaluate the improved usability of the proposed infrastructure via case studies and user studies.

Configuration scripts are used to manage system configurations and provision infrastructure at scale. Configuration scripts are susceptible of including security weaknesses such as hard-coded passwords, which can facilitate large-scale data breaches, as well as provisioned systems being compromised. We propose an automated technique to identify security weaknesses so that configuration scripts do not cause large-scale security attacks and data breaches. We will build upon our recent research and construct eSLIC, which will overcome previous limitations of our initial prototype and facilitate wide-spread security static analysis of infrastructure. We will make eSLIC available for OSS and practitioners in industry.

Recent years have seen a surge in popularity in smart home IoT products, and with the ongoing pandemic, people are spending more time interacting with such devices. However, it is unclear whether and how these IoT devices affect the security, privacy, and performance of the home network as well as the access network. In this proposal, we focus on developing privacy-preserving IoT analytics to help network providers allocate network resources accordingly and, at the same time, help consumers identify potential anomalous behavior.

This proposal introduces innovative strategies to fortify the metaverse ecosystem, aiming for secure, resilient, and privacy-enhanced digital world experiences. The proposed work will tackle physical, digital, and virtual security and privacy risks, developing reliable digital mapping and robust ML services, as well as privacy-enhancing user interactions.

We propose AERPAW: Aerial Experimentation and Research Platform for Advanced Wireless, a first-of-its-kind aerial wireless experimentation platform to be developed in close partnership between NCSU, Wireless Research Center of North Carolina (WRCNC), Mississippi State University (MSU), University of South Carolina (USC), City of Raleigh, Town of Cary, Town of Holly Springs, North Carolina Department of Transportation (NCDOT), and numerous other project partners. With a major focus being on aerial communications within low altitude airspace, AERPAW will develop a software defined, reproducible, and open-access advanced wireless platform with experimentation features spanning 5G technologies and beyond. NCSU, USC, and MSU researchers have existing UAS experimentation capabilities and ongoing experimental research activities involving wireless technologies spanning software defined radios (SDRs), LTE, WiFi, ultra-wideband (UWB), IoT, and millimeter wave (mmWave), which will form the initial baseline framework for the AERPAW platform. To deploy AERPAW, NCSU will work closely with NCDOT�s Integration Pilot Program, a three-year FAA project that allows BVLOS UAS experimentation for medical supply delivery in North Carolina, in close collaboration with NCSU, several UAS companies, municipalities, and a medical institution. Initial flight tests have already started within the Raleigh area, and will be expanding to other parts of the state in 2019 and beyond. Any additional FAA permits, as necessary, will be secured by AERPAW team in close collaboration with NCDOT.

The global cellular telecommunication system is critical infrastructure that has become a ubiquitous platform for Internet connectivity supporting a wide range of use cases for both consumers and industry. We are now on the cusp of widespread adoption of 5G technology. While 5G is widely marketed for its gigabit per second rates and ultra-low latency, it also also fundamentally changes the internal network architecture, providing dynamic provisioning of software-defined services that offer enhanced control to network tenants including virtual operators and enterprises. This new threat model necessitates deep investigation of the many technical components that comprise the cellular system. Whereas several initial studies have formally modeled and evaluated the security of 5G cryptographic protocols, little is known about the security of software and hardware systems that implement them. To this end, the goal of this work is to aid mobile network operators in deploying secure cellular systems through the development of tools and techniques that extract, model, and analyze security-sensitive logic of the source and binary code that exists within cellular system functional entities.

The DoD makes available scholarships for NC State students specializing in cybersecurity. Students (junior and senior) must be US citizens and majoring in any engineering field at the bachelor, master's, and PhD level with a specialization in cybersecurity. In addition to full tuition, this scholarship can last up to five years and provide a generous stipend, tuition, health insurance, and an allowance for other professional expenses. In return, the student agrees to work after graduation with a federal, executive-branch government agency for an equal period of time. The program includes mentoring, professional opportunities while in school, and assistance finding internships and post-graduation full-time employment in government.

The DoD is making available scholarships for N.C. State students specializing in cybersecurity. Students (junior and senior) must be US citizens and majoring in any engineering field at the bachelor, master's, and PhD level with a specialization in cybersecurity. In addition to full tuition, this scholarship can last up to five years and provide a generous stipend, tuition, health insurance, and an allowance for other professional expenses. In return, the student agrees to work after graduation with a federal, executive-branch government agency for an equal period of time. The program includes mentoring, professional opportunities while in school, and assistance finding internships and post-graduation full-time employment in government.

This project aims to develop an integrated lightweight and energy-efficient prosthetic care robot framework. It will enable proactive and user-specific prosthetic control to improve walking function in a variety of walking conditions found ubiquitously in daily living.

The expansion of K-12 Computing Education Research (CER) is quickly following the expansion of computing education in primary and secondary schools, yet much remains to be learned about the effectiveness of the implementation and the quality of evidence produced by research. While the integration of computing education into K-12 in the United States is still in its infancy, so is the research necessary for identifying promising practices for educational outcomes across a variety of populations, including those historically underserved and marginalized by education. The proposal seeks to 1) summarize and frame prior equitable K-12 CS education research against the areas of capacity, access, participation, and experience as defined by the CAPE framework; 2) develop publicly-available recommendations and resources along the CAPE framework for expanding coverage of equitable K-12 computing education research; and 3) design and pilot workshops to train K-12 education research in methods and practice to support robust evidence-based research results that can inform practice.

In this research, we plan to characterize the help resources available to students in Computer Science (CS) courses and analyze the order and frequency of use by the students. We will also study why students choose specific help patterns and what help they perceive to be effective for their learning. Our goals are to explicitly teach students about the help resource landscape, guide them to identify CS topics where they may need help and to empower them to be more effective in traversing the complex landscape of help resources.

Educating the next generation of cybersecurity professionals is a critical need for the State of North Carolina and the United States. We are utilizing our expertise in cybersecurity research to prepare undergraduate and Masters computer science students at NC State for cybersecurity jobs. Scholarship for Service (SFS) will provide students from North Carolina and the United States, especially from underrepresented groups, the opportunity to receive a high quality cybersecurity focused degree. SFS students will be part of a larger cohort of cybersecurity students who will participate in supplemental activities, events, and conferences as part of their educational experience.

This project aims to address these financial and computational challenges by developing innovative performance measurement and analysis techniques tailored for deep learning workloads and encapsulating these techniques into a profiling toolkit (i.e., DLToolkit). Building upon existing open-source profilers, DLToolkit will offer scalable analysis, aggregation, and visualization of deep learning workloads, providing invaluable insights to scientists, thus fostering expedited innovation in scientific applications using deep learning.

Learning how to code is a key and challenging component in computer science (CS) education. Traditionally, the primary focus in programming courses has been on achieving functional correctness. However, there is a growing recognition of the importance of program performance, as evidenced by factors such as execution time, memory usage, and other metrics. This shift has garnered attention from both students and instructors, highlighting the need to incorporate performance considerations alongside functional correctness in CS education. This project will develop EduPerf, which aims to hoist program performance as the first-order metric in CS education via tightly integrating program performance analysis into different levels of CS courses for both students and instructors.

Continuous Integration (CI) has become an essential component of the modern software development cycle. Developers engineer CI scripts, commonly called workflows or pipelines, to automate most software maintenance tasks, such as testing and deployment. Security issues in workflows can have devastating effects resulting in supply-chain attacks. We propose to handle these research challenges by (1) defining a threat model and deriving security properties from first principles; (2) developing a framework based on our Workflow Intermediate Representation (WIR) that enables us to verify and define security properties in a platform-agnostic way.

We study the web differently from how users explore it, as browsers are not meant to be monitoring tools. Researchers build either ad-hoc solutions or use high-level information from the browser that is inadequate to identify some of the most advanced web attacks. This research aims at building the fundamental blocks for studying an increasingly complex web by developing a monitoring platform that sheds light into the inner workings of modern browsers and websites. Our research outcomes will allow any researcher, web developer or web user to understand better how the web works.

This research proposes to develop a comprehensive framework for evaluating the offensive capabilities of Large Language Models (LLMs) in understanding, analyzing, and potentially exploiting newly published Common Vulnerabilities and Exposures (CVEs) in real-time. The study aims to establish standardized metrics for assessing LLMs' offensive security capabilities and quantifying how LLMs are improving offensively over time.

Fingerprinting has been a known threat to web privacy for over a decade. Yet, automated detection of fingerprinting methods and scripts has been lacking the properties for protecting web users from such an evolving web threat. Our proposed work aims to provide novel detection methods for browser fingerprinting both at its core, the browser and the evolution of its APIs, and at the page level, via dynamic analysis ofJavaScript. We also propose developing countermeasures that are capable of performing more fine-grained blocking not only at the script level, but also at the API level where an instance of a script/API will be blocked depending on inferring the underlying intent behind executing the script or accessing the API.

Artificial Intelligence (AI) has emerged as a foundational technology that is profoundly reshaping society. With accelerating advances in a wide array of capabilities including natural language processing, computer vision, and machine learning, AI is quickly finding broad applications in every sector of society. Critically, AI holds significant transformative potential for improving human learning. This National Artificial Intelligence (AI) Research Institutes proposal centers on the establishment of the Institute for an AI-Engaged Future of Learning. Driven by a learner-centered vision of the potential of AI-augmented learning, the ENGAGE AI Institute will conduct (1) foundational AI research on natural language technologies, computer vision, and machine learning and (2) use-inspired AI research on AI-augmented learning, thereby creating learning experiences specifically designed to promote student engagement in formal and informal learning settings. The ENGAGE AI Institute brings together an exceptional interdisciplinary team spanning five organizations with deep expertise in AI and education, including four universities (North Carolina State University, the University of North Carolina at Chapel Hill, Vanderbilt University, and Indiana University) and Digital Promise, which will serve a �nexus� role for the Institute. The Institute will create AI-augmented learning technologies with specific foci on supporting two forms of engaging collaborative inquiry learning experiences: collaborative learning (problem solving and learning that play out in groups) and embodied learning (learning processes that are grounded in the interplay between the body, movement, and senses). The Institute will focus on AI-driven narrative-centered learning environments that create engaging story-based problem-solving experiences to support collaborative inquiry learning. The Institute will explore AI-augmented learning that operates at three levels: individuals, small groups, and larger groups within a range of educational contexts (e.g., classrooms, museums).

The U.S. Army???s Force Modernization strategy highlights the critical role synthetic training will play in transforming Soldiers to operate as a multiple domain force. A key affordance of synthetic training environments is their capacity to support competency-based experiential learning (CBEL), which prescribes an active approach to learning and expertise development that incorporates adaptive instruction and intelligent tutoring capabilities. Although synthetic training environments show great promise for supporting CBEL, there is a lack of guidance on how synthetic training experiences should be integrated into Army schoolhouse curricula to support competency development and experiential exposure. To maximize the effectiveness of CBEL, synthetic learning experiences need to be dynamically crafted to support individual learning needs and skill development. Simulation-based training scenarios can offer trainees valuable experiences but are resource-intensive to create, and in most cases, schoolhouses have a limited supply of scenarios that they can utilize for a particular course. Competency-based scenario generators offer considerable promise for addressing these challenges by tailoring synthetic training experiences to the needs of individual learners in support of CBEL. Competency-based scenario generators can dynamically shape training experiences, scenario events, unit behaviors and states, and virtual environments in order to support CBEL. Scenario generators can leverage recent advances in machine learning to provide data-driven approaches to support competency-driven training. Recognizing the opportunity introduced by recent advances in machine learning and data-driven scenario generation, the proposed project will investigate how we can devise generalizable, data-driven scenario generation models that dynamically generate training scenarios that achieve target learning objectives to support CBEL in Army schoolhouses. We will design and develop the CompGen competency-based scenario generation framework and demonstrate its data-driven capabilities for supporting CBEL in an institutional training setting.

It has long been recognized that drawing can be a powerful approach to learning. Learning-by-drawing activates a complex set of cognitive processes that requires students to deeply engage with a subject matter. The project centers on the design, development, iterative refinement, and investigation of a sketch-based science learning environment. Specifically, the project will focus on the development and piloting of a sketch-based science learning environment to support students� conceptual understanding of science with an emphasis on modeling. The project will culminate in a pilot study to investigate the effectiveness of the sketch-based learning environment for improving students� factual understanding, their inferential understanding, and their ability to engage in science modeling. By utilizing a mixed methods approach integrating quantitative and qualitative work with learning analytics, it is anticipated that the project will yield theoretically-driven, empirically-based advances in sketch-based science learning environments that significantly improve conceptual understanding of science in upper elementary students.

Developing our biomedical workforce is a critical national need. Artificial intelligence (AI) has emerged as a powerful technology that will play an important role in biomedical careers. The project will engage middle school students learning about AI in the context of biomedical careers through the design, development, and evaluation of AI4Health, a game-based learning environment that will create personalized adventures in which students will utilize AI tools to solve biomedical problems. The project will evaluate the impact of AI4Health on students??? knowledge, interest, and self-efficacy for AI and biomedical careers.

Large language model (LLM) systems, including retrieval augmented generation (RAG) systems, agents, and broad LLM-integrated applications, are widely deployed in the real world. LLM systems are susceptible to a variety of security and safety threats, raising concerns about their reliability and trustworthiness in critical applications. The goal of this proposal is to bridge the gap between general cybersecurity education and LLM infrastructure research and train a group of LLM research workforce with advanced CI cybersecurity knowledge to undertake specific roles in their future careers.

As the correlation of data gains importance in many domains, high-dimensional tensors are becoming an ever more important object to represent data and analyze its inherit properties. Tensor networks targeting very high-dimensional data and extracting physically meaningful latent variables are underdeveloped because of their complicated mathematical nature, extremely high computational complexity, and more domain-dependent challenges. This work proposes Cross-layer cooRdination and Optimization for Scalable and Sparse Tensor Networks.

The aim of this project is to achieve early, rapid, and precise detection of harmful downy mildew on cucurbit plants to enhance crop health and production. This objective will be accomplished by employing quantum sensing-enabled spectroscopy, which utilizes entangled photons and a quantum machine learning receiver. The resulting quantum sensing device will be incorporated into a robotic land rover for testing in North Carolina's cucurbit fields.

This project will create a general-purpose, open-access, and programmable quantum network prototype for the quantum information science and engineering (QISE) community to experiment new quantum technologies and train teachers and students. The key applied methodology is virtualization that permits rapid and flexible experimentation via agile software controls, without resorting to daunting hardware modifications. The research team initiates a research agenda consisting of three thrusts, namely re-designing key quantum components, developing communication protocols, and implementing the prototype in the testbed. A new curriculum based on this prototype will be created and disseminated to train a large body of college students.

Wireless devices are inherently faculty which can result in multifaceted data errors in computing, caching, and communications (C3). These errors have been widely deemed harmful, but recent studies have shown that they can be benign or even beneficial. The research objective of this project is to proactively harvest, render, and control data errors across C3 of wireless devices for significant performance gains in energy efficiency, throughput, data privacy, etc. Moreover, the research efforts will be coupled with educational innovations through the development of new laboratories, lecture contents, outreach demos and a novel undergraduate/graduate co-learning pedagogy.

Recently, research progress has been made in robust machine learning (ML) to understand and enhance the robustness of ML in close-world environments, where the training and testing settings are knowable and strictly controlled. However, substantial challenges remain in generalizing this robustness to realistic open-world environments, where crucial factors such as data distributions and threat models may be unknown, variable, or uncontrollable. The overarching research goal of this project is to close the enormous research gap between robust ML in close-world environments and open-world environments. We propose three novel research thrusts to pursue generalizable, configurable, and certifiable robust ML.

The project presents a rigorous framework for resilience-native wireless enabled by hybrid digital-physical intelligence, utilizing a combination of network optimization, graph theory, machine learning, experimental measurements and models that mix the physical and virtual contexts. This project aims to overcome the challenges of incorporating the fundamental predictive power, wireless twinning, and topological mobility capability into resilience design with new models, theory, and systems.

Digital twin is emerging as a revolutionary approach to testing and assurance for next-generation (nextG) wireless networks enabling continuous prototyping, optimization, and validation. The primary goal is to lay the foundations of digital network twin (DNT) by exploring innovative technologies to map and optimize nextG wireless networks in twins, thereby facilitating development, testing, and formal evaluation exercises of nextG wireless networks. The research agenda comprises two thrusts. Thrust 1 is focused on novel approaches of building the twin environment to replicate the physical network world. Thrust 2 shall build and optimize the network twins over actual network environments associated with communication, computing, and networking resources. The fundamental research of Thrusts 1 and 2 is then implemented in a developed DNT platform used to demonstrate the behavior and performance of designed twining and optimization approaches.

This collaborative project between NCSU and ETS is focused on developing new noninvasive process-based measurements for students engagement with writing tasks, including analyses of their writing quality, working habits, and responses to feedback. As part of this project we will develop a secure instrumented platform for online writing tasks that will provide analytical tools for instructors and researchers to monitor and evaluate student's work.

Recent years have seen a growing recognition of the national STEM workforce shortage. Although problems abound in all STEM disciplines, the shortage is particularly acute in information and communications technology. This is especially true in artificial intelligence (AI), a field of computer science that focuses on the design of computing systems that solve problems involving human-like capabilities including reasoning, learning, and natural language. Engaging middle-grade students, especially those from underserved populations, in artificial intelligence through the creation of lifelike AI for digital games offers a promising approach to encouraging students to pursue innovative computing careers. The AI Play project will engage students in a broad range of computing activities centered on creating AI for games. The project will see the development of a learning environment and curriculum that introduces artificial intelligence into middle school emphasizing connections to the CSTA K-12 Computer Science Standards. The AI Play project will host a series of five-day camps for underserved populations where students will engage in hands-on learning activities under the guidance of teachers and undergraduate computer scientists, who will serve as mentors and role models as the students engage in artificial intelligence, while designing and developing AI for games. The final year of the project will see an evaluation of the AI Play program and its impact on students� learning and interest in artificial intelligence.

The objective of this work is to (1) develop sample programs that utilize the SCR library and can serve as benchmark examples to the community as well as (2) devise novel methodologies for improving the performance of checkpoint/restart on modern HPC systems with an implementation and evaluation.

The primary goal of the QACTI quantum system and technology demonstrator is to build an advantage-class trapped-ion quantum-computer capable of being used by the broader scientific community remotely. The secondary goals are to discover algorithms suited for near-term quantum computers, improve and democratize ion trap quantum technology, and develop a workforce capable of utilizing and building advantage-class machines.These goals only become achievable by performing device-oriented experiments at fine-grained control levels that are not available on commercial platforms, thereby contributing to the development of a combined hardware/software stack in an open-source manner.

This project proposes to explore solutions to both the workflow scheduling problem and the fault awareness requirements. Its primary aim is to prototype a novel software scheduling technology to manage dynamically changing heterogeneous resource pools on one side and a fault propagation mechanism within Flux instances on the other side. This reflects emerging trends in combing HPC and cloud computing while taking full advantage of characteristics for heterogeneous resource requirements of modern applications workflows under challenges to address faults in a transparent manner.

The Institute for Robust Quantum Simulation will focus on using quantum simulation to gain insight into�and thereby exploit�the rich behavior of complex quantum systems. Combining expertise from researchers in computer science, engineering, and physics, our team will address the challenge of robustly simulating classically intractable quantum systems of practical interest, and verifying the correctness of the simulation result.

Open-source robot software aims to enable rapid system development but comes with little or no tooling for automated testing and analysis. This work utilizes model checking of behavior trees and abstract type inference of physical units to automatically suggest system tests and to help ensure the absence of certain classes of software defects. This CAREER proposal examines whole system representation and tooling across interdisciplinary boundaries. We aim to substantially reduce the cost and improve the scalability of lightweight formal methods for robotic software systems, thus laying the foundation for the next generation of automated testing and analysis of robotic systems.

Machine learning (ML) is a powerful computing tool for building models from data, which is becoming a vital skill across STEM disciplines. However, ML is a challenging subject, requiring students to construct complex ML "pipelines," often with little one-on-one support from instructors. The goal of this CAREER proposal is to aid students in learning to design and implement ML pipelines through a data-driven tutoring system. To do so, the project will develop novel techniques for evidence-centered, real-time assessment of students' ML knowledge and novel forms of automated support for ML, including design feedback, and adaptive code examples.

The goal of this work is to investigate the role of self-regulated learning (SRL) in computing education by validating and analyzing fine-grained trace data from students' interactions with programming tools. We will: 1) Conduct instructor interviews and classroom observations to identify SRL strategies related to programming tool use; 2) Instrument the tools to record student behavior, adding a priori design choices that make students' SRL strategies more visible; 3) Conduct laboratory studies and collect think-aloud protocols, then code the data with strategies identified earlier; 4) develop educational data mining techniques to identify SRL behaviors from log data; 5) deploy the SRL detectors in both introductory and more advanced CS classrooms, using the detected behaviors to validate and extend SRL theories in the domain of CS.

We propose to develop infrastructure to enhance and scale CSEd research by leveraging the power of data-driven AI and ML. To do so, we need to overcome 3 challenges: data (there is not enough quantity and quality of data), analytics (developing and sharing data mining and AI methods for CSEd is highly siloed and disconnected) and evaluation (AI-based interventions and tools are not easily deployed and replicated). To address these challenges, we will develop a large collection of resources including datasets, analytical approaches, reusable smart learning content, and tools and user services that enables the community to reuse the resources and contribute to the collection.

We will develop new data-driven methods to support students automatically as they create novel, open-ended and creative, computational artifacts. Specifically, we will develop techniques to adaptively scaffold project design and planning, detect students' programming goals, offer on-demand example-based support and tailor help to students needs through an interactive help interface. We will augment the popular Snap programming environment, which is already used in hundreds of high school and college classrooms, with these features and evaluate their effective in a series of experiments designed to explore how students approach open-ended tasks and how best to support them.

Optimal transport (OT) is a powerful tool for comparing probability distributions and computing maps between them. OT has been studied extensively in mathematics, engineering, physics, economics, operations research, and computer science because of their numerous applications. Despite extensive work, computing OT plans has remained a computationally challenging problem, and there is a large gap between the theory and practice of OT algorithms. In this project, we bridge this gap by designing and implementing novel algorithms for computing exact and approximate OT as well as for data analysis on a set of distributions with OT as a metric.

Telephone users are regularly besieged by unsolicited sales and scam calls, cannot verify identities of callers, and enterprises frequently fall prey to expensive compromises of their telephone infrastructure. This proposal will deliver techniques to detect these issues, conduct network-wide systematic measurement, and provide practical defenses for these problems. The vision of this 5-year project is to provide technologies that will restore the telephone network to its former status as a trusted and trustworthy network.

Animal-Assisted Interventions (AAIs) are goal-oriented programs that intentionally incorporate animals, such as dogs, for therapeutic benefits. AAIs are widely used in a variety of settings, including for cancer patients and veterans with Post-Traumatic Stress Disorder (PTSD). AAI has proven to provide physiological, psychological, and symptom benefits. The positive effects of AAI are posited to be, in part, due to the dynamic human-animal bond (HAB). Despite AAI???s popularity, neither comprehensive AAI assessment methods nor stakeholder-informed, standardized AAI protocols exist???activities critical for understanding AAI mechanisms of action, the role of the HAB, and the mechanisms by which the HAB is formed and maintained. The overall objective is to develop, test, and evaluate an IoT software and hardware system for dyadic physiological monitoring of humans and animals in AAI settings, and to innovate in analytic methods for interpreting the data. The work is important, careful, and systematic and will yield novel capabilities and information for AAI outcome assessment and intervention development. Our team has developed and tested an innovative platform that incorporates wearable, wireless sensors that will simultaneously gather physiological data (i.e., activity, heart rate/variability, respiratory rate, electrodermal activity) from both humans and dogs involved in AAIs. This novel system will be combined with psychological (i.e., distress, well- being) and symptom data (i.e., pain, fatigue) collected from the patient and dog handler. Tasks will include qualitative methods (i.e., focus groups) to elucidate first stakeholder-informed, standardized AAI protocols. Conducting focus groups including patients, handlers, and providers will provide information to optimize a structured AAI protocol that can be delivered with a high level of intervention fidelity, lead to beneficial patient outcomes, and provide controlled settings for objective, continuous measurement of both patient and dog physiology and behavior.

The Agricultural DECision Intelligence moDEling System for huMan-AI collaboRative acTion Elicitation and impRovement (DECIDE-SMARTER) project will lay the foundations of democratized access to Decision Intelligence (DI) technology for stakeholders across the agriculture value chain, filling a longstanding gap between technology and decision makers. Through a process of participatory design, the project team will work with stakeholders in the sweetpotato value chain to: 1) Create a software asset that helps growers with an otherwise difficult decision; 2) conduct experiments that inform the best software interfaces possible to support complex agricultural decision making (through characterizing, understanding, and leveraging human cognitive abilities; 3) identify potential sources of bias in the DI process that would present barriers to democratized access to the technology; and 4) develop a reference architecture and prototype implementation of a modeling, simulation, and visualization framework for implementing multiple DI models with agriculture stakeholders. The project will leverage the ongoing research, data acquisition, and stakeholder efforts by the Sweetpotato Analytics for Produce Provenance and Scanning (Sweet-APPS) team, a multi-disciplinary endeavor that aims to reduce agricultural waste and maximize yield for North Carolina’s sweet potato growers.

NSF Computer and Information Science and Engineering Graduate Fellowship Award (CSGrad4US) for incoming CSC PhD student Jordan Esiason.

This proposal will generate novel abstractions for computing that extend serverless functions to better leverage unique hardware characteristics, and for memory to allow more automated leveraging of workload characteristics such as locality and compute intensity. Further, this work expands currently limited secure enclaves to include parallel, heterogeneous hardware needed to support a wide range of applications, and enhances serverless databases to leverage heterogeneous compute resources.

Memory safety is essential. Despite decades of research, unauthorized memory reads and writes are still among the most common security attacks. The emerging persistent memory (PM) amplifies the importance of strong memory protections. As a promising supplement or substitute of DRAM as main memory, PM offers higher density, better scaling potential, lower idle power, and non-volatility, while retaining byte addressability and random accessibility. Data in a PMO is long lived; its existence and structure are preserved across process runs. The longevity, plus direct byte-addressability, makes it more vulnerable as attacks to a PMO could span across executions. This proposal aims to improve the understanding of the problem and provide innovative solutions to strengthen memory security for future NVM-based systems.

This project will develop IRIS-D: a data flow-enabled portable memory abstraction for seamlessly orchestrating memory in diverse heterogeneity. Using the data-flow analysis, IRIS-D will guard against race conditions while multiple heterogeneous devices can access memory objects and optimize data movement both between the host and devices and between devices across multiple nodes in the distributed multi-node systems. As a result, the proposed IRIS-D will provide high programming productivity, performance, and portability for distributed multi-device heterogeneous execution in HPC and cloud systems with diverse co-existence of architectures from different vendors, including CPUs, NVIDIA GPUs, AMD GPUs, field-programmable gate arrays (FPGAs), and Hexagon digital signal processors (DSPs).

Stroke is a leading cause of motor disability. A majority of stroke survivors exhibit upper and lower limb motor impairments, ranging from incapability of reaching and grasping objects to limited ambulation. The objective of this project is to develop a personalized, community-based rehabilitation system to improve daily functions of stroke survivors. The system will include three essential components – a nanomaterial-enabled multifunctional wearable sensor network to monitor arm and leg functional activities; a low-power data acquisition, processing, and transmission protocol; and a user interface (i.e., smart phone APP) to communicate training outcomes to the users and clinicians and receive feedback from the users and clinicians. The proposed community-based rehabilitation system will enable personalized, continuous rehabilitation during daily activities.

Data reduction holds paramount importance in scientific endeavors and various data-intensive domains. This necessity is compounded by the unprecedented surge in data volume propelled by advancements in facilities and scientific research. Data compression stands as a prevalent method for data reduction. The prevailing solutions are hindered by a fundamental constraint: the requirement for decompression prior to processing. This project endeavors to overcome the fundamental limitation by developing novel methods to enable efficient processing on compressed data directly.

Coastal areas are highly susceptible to significant flood damage from sea level rise, high tides, storm surges, and extreme rainfall due to dense populations, high property values, and disproportionately vulnerable populations in low-lying areas. Understanding and predicting the consequences of stresses and shocks in this coupled land-ocean system is vital for the future viability and sustainability of the region. Researchers, encompassing faculty, postdocs, graduate, and undergraduate students, are eager to engage in Environmental Science (ES) research, a critical endeavor to safeguard regions against environmental disasters. However, the ES data collected from distributed sensors with various monitoring resolutions worldwide, both diverse and complex, presents challenges but also opportunities for scientists engaged in understanding and predicting environmental phenomena using numerical modeling and/or artificial intelligence (AI) algorithms. To this end, a widely accessible AI research cyberinfrastructure (CI) that brings together powerful computational resources, data, testbeds, algorithms, software, services, networks, and expertise is important and helps democratize the AI research landscape and enable more efficient ES research. Recognizing this, our project aims to train the next generation of CI professionals and contributors in the two universities located in the coastal states: Florida International University (FIU) and North Carolina State University (NCSU). Our long-term goal is to build a sustainable and transdisciplinary CI community to support the nation???s advanced CIs that can ensure broad adoption of advanced CI resources and expert services including platforms, tools, methods, software, data, and networks for research communities, to catalyze foundational AI research advances, and to enhance researchers' abilities to lead the development of new CI through education, training, and workforce development.

Recent advances in artificial intelligence have raised concerns of ethics in regards to intelligent, adaptive agents. This project begins from a model of a sociotechnical system (STS) comprising autonomous social entities (people and organizations -- principals) and technical entities (agents, who help principals). Its objective is to uncover principles of multiagent systems that enable developing sociotechnical systems that incorporate ethical concerns adaptively and from multiple perspectives, and with high confidence. This project will develop a formal computational representation of an STS in terms of social controls over its principals and technical controls over its agents.

Semantic code search uses behavioral specifications, such as input/output examples, to identify code in a repository that matches the specification. Challenges include handling scenarios when 1) there are too few solutions, 2) it is difficult to understand how solutions differ, and 3) there are too many solutions. I propose techniques to 1) expand the scope of code that can be modeled and find approximate solutions when an exact one does not exist, 2) determine the differences between two code fragments, and 3) navigate a large space of possible solutions are needed by selecting inputs that maximally divide the solution space.

This project advances the state of knowledge about how to infer misconceptions and generate explanations without any explicit models of a programming language. In contrast to existing approaches, which involves manual identification of misconceptions in programming languages, or cross- language migrations�which provide translations but no explanations�our technique automatically discovers inconsistencies cross-languages and supports automatic resolution for problematic translations.

Software testing is a critical skill for computer science graduates entering technical positions. Software tests, and in particular unit tests, have several uses in education. The purpose of this proposal is to create pedagogy and tools around writing unit tests for CS3 and Software Engineering (SE) courses. Building on our preliminary work, we develop and evaluate the impact of a lightweight intervention with explicit testing strategies on the test quality of student-written tests. Then, we investigate the impact of the process of writing tests on student outcomes.

This project aims to address the following question: How to model the combined information of a pan-genome collection succinctly (and in a biologically meaningful way) such that the genomic analysis on that representation is both easy-to-compute and accurate? Pan-genome collections may be represented as high-scoring Multiple Sequence Alignment (MSA) data, indexed text data, or the more popular graph-based representations (pan-genome graphs). These models need to support read mapping queries efficiently. This research will lead to a new class of string/graph algorithms for the analysis of pan-genomic data.

This Research-Practice Partnership, AI by 8, seeks to empower Kindergarten through 2nd grade teachers with pedagogical content knowledge and methodologies to introduce the fundamentals of AI through engaging, unplugged activities (requiring no devices) integrated into their language arts instruction. These AI-infused language arts lessons will be co-designed with a core set of 10 AI Fellows and implemented with an additional set of 20 AI Implementers, resulting in potential impact for more than 600 rural, K-2 students.

The proposed project will see the design, development, and investigation of a multimodal affect-sensitive learning environment for generating student interest in middle school science. We will capture rich multi-channel data (eye gaze, facial expression, posture, interaction traces) on student problem solving with an inquiry learning environment. We will utilize multimodal machine learning to induce affect recognition models, which will drive run-time affect-sensitive interventions to trigger and sustain student interest. The project will culminate in a classroom experiment to evaluate the impact of the multimodal affect-sensitive learning environment on student learning and science interest.

Fueled by machine learning (ML) model and hardware advancements, intelligence is transforming every walk of life. For critical applications like autonomous vehicles, ensuring inference dependability is essential. Unfortunately, current hardware cannot provide such a promise. This CAREER project aims to create a new paradigm of safeguarding ML execution against both passive hardware faults and active fault attacks. The novelties lie in the new capability development inside ML processing, and the cross-layer exploration of algorithm, architecture, and hardware security. The broader impacts include yielding practical solutions for ensuring the root of trust of accelerated intelligence services and abundant educational opportunities.

Machine learning (ML) as a service on cloud is pervasive, but it poses real threats to personal or business providers' privacy. To guarantee privacy, cryptographic protocols, such as Homomorphic Encryption, are promising due to enabling ML analytics directly on encrypted data. However, there exists a big gap between the theory and practice, e.g., long latency due to the prohibitively expensive computation or communication overhead over ciphertext. This project aims to practically accelerate the private ML service by offering a full-fledged development of efficient, scalable and encryption-conscious computing paradigms. The broader impacts include advance trustworthy artificial intelligence and educational opportunities.

Future computer data centers face increasing demands for high-computation workloads driven by power-hungry deep neural network (DNN) models. DNN accelerators, employing processing in memory with new storage devices, promise energy efficiency and performance. However, these accelerators encounter stability challenges due to device limitations. This project aims to develop efficient neural network approaches to address this issue. The impact includes potent, scalable deep learning systems benefiting various applications and fields, such as business, science, and national security. It also elevates students' competence and confidence in the competitive job market, integrating research results into courses and outreach activities.

Modern distributed systems and Internet services require authentication between their components to protect their services from unauthorized access and ensure appropriate billing. In practice, this authentication is performed by presenting a static secret, such as an �API key� or password. These are difficult for developers to manage and deploy securely, and credentials are accidentally or intentionally stored in widely readable software repositories. This threatens not just the security of the leaker, but also the authenticating service. The ultimate root cause of this issue is the adaptation of user authentication methods (e.g., passwords) to software in ways that are inappropriate and ultimately unsafe. This proposal will fund research to more reliably and consistently identify these leaked software credentials, triage them according to the risk they present, conduct developer interventions to train them to properly manage this risk, and finally develop more secure yet manageable alternative solutions to software authentication.

This proposal aims to develop techniques that enable application of robust predictive intelligence algorithms in the new spectrum era. The goal is to ensure robustness of predictive intelligence when handling critical spectrum-related tasks, including but not limited to: spectrum management, spectrum trading and spectrum monitoring.

This project aims to investigate the possibility and develop the technical foundation of building an open, decentralized wireless access ecosystem. The core contribution is around building contract overlay networks to enable on-demand spectrum leasing and wireless access, enabling verifiable contract fulfillment, and incentivizing broad and honest participation in the ecosystem.

This project seeks to develop theoretical tools (models and algorithms) for analyzing and optimizing a hybrid continuous-discrete variable quantum network architecture for the future quantum internet.

Abstract: The goal of this CAREER project is to fill the gap between growing application complexity and performance requirements, and existing application-agnostic network management, to enable and incentivize rigorous performance guarantees for distributed real-time applications at the network edge. The core contribution is the design, analysis, and evaluation of WolfPack, a general edge resource provisioning framework for real-time applications. The PI will focus on three key thrusts: 1) modeling and optimization of edge resource provisioning, 2) stochastic models and robustness techniques to control the risk, and 3) incentive mechanisms to enable truthful and competitive network edge resource trading.