BJC-STARS is a CS10K proposal to broaden access to Computer Science education through engaging colleges and universities to prepare and support regional communities of high school teachers to teach the Beauty and Joy of Computing (BJC) Computer Science Principles course. We will leverage the successful STARS model focusing on engaging faculty and students in a community committed to leading efforts to broaden participation in computing. Each year, we will engage new university faculty who will teach BJC and facilitate professional development and support to high school teachers and students. We will also build a STARS community among participating high school teachers and students, engaging them in the need to broaden participation in computing.

In this FRABJOUS CS project, we will prepare 60 high school teachers to teach the Beauty and Joy of Computing (BJC) Computer Science Principles curriculum. The BJC course is a rigorous introductory computing course that highlights the meaning and applications of computing, while introducing low-threshold programming languages Snap-Scratch, GameMaker and AppInventor. BJC is informed and inspired by the Exploring Computer Science curriculum, that was explicitly designed to channel the interests of urban HS students with ?culturally relevant and meaningful curriculum? [Goode 2011][Margolis 2008]. The BJC course uses collaborative classroom methods including pair learning, and student-selected projects are geared toward leveraging students? knowledge of social media, games, devices, and the internet. At UNC Charlotte in 2010 and 2011, PI Barnes engaged college students in supporting the BJC course, and in after-school outreach and summer camps that excite middle and high school students about this curriculum at different levels. The project engages three university faculty members and 6 college students to help the high school teachers build a Computer Science Teachers Association chapter and provide ongoing professional development and support for the BJC course. The project also engages high school teachers and an education researcher to help refine and enriches the BJC curriculum to be easier to adopt and teach in high schools.

In this FRABJOUS CS project, we will prepare 60 high school teachers to teach the Beauty and Joy of Computing (BJC) Computer Science Principles curriculum. The BJC course is a rigorous introductory computing course that highlights the meaning and applications of computing, while introducing low-threshold programming languages Snap-Scratch, GameMaker and AppInventor. BJC is informed and inspired by the Exploring Computer Science curriculum, that was explicitly designed to channel the interests of urban HS students with ?culturally relevant and meaningful curriculum? [Goode 2011][Margolis 2008]. The BJC course uses collaborative classroom methods including pair learning, and student-selected projects are geared toward leveraging students? knowledge of social media, games, devices, and the internet. At UNC Charlotte in 2010 and 2011, PI Barnes engaged college students in supporting the BJC course, and in after-school outreach and summer camps that excite middle and high school students about this curriculum at different levels. The project engages three university faculty members and 6 college students to help the high school teachers build a Computer Science Teachers Association chapter and provide ongoing professional development and support for the BJC course. The project also engages high school teachers and an education researcher to help refine and enriches the BJC curriculum to be easier to adopt and teach in high schools.

The REU Site at NC State University will immerse a diverse group of undergraduates in a vibrant research community of faculty and graduate students working on cutting-edge games, intelligent tutors, and mobile applications. We will recruit students from underrepresented groups and colleges and universities with limited research opportunities through the STARS Alliance, an NSF-funded national consortium of institutions dedicated to broadening participation in computing. Using the Affinity Research Groups and STARS Training for REUs models, we will engage faculty and graduate student mentors with undergraduates to create a supportive culture of collaboration while promoting individual contributions to research through just-in-time training for both mentors and students throughout the summer.

Despite long-standing awareness that social interaction is an integral part of knowledge construction, efforts to study complex collaborative learning have traditionally been relegated to qualitative and small-scale methodologies. Relatively new data traces left by online learning environments, including massive open online courses (MOOCs), offer the first real hope for scaling up such analyses. The purpose of the proposed research is to develop comprehensive models for collaborative learning which in turn will enable instructional design and the authentic assessment of the individual within the group context. This task is undertaken by an interdisciplinary team of researchers with specializations in natural language processing, discourse analysis, social network analysis, educational data mining and psychometrics.

The Beauty and Joy of Computing project presents a unique opportunity to scale the STARS Alliance further while also enhancing national efforts to engage more high school teachers and students in teaching and learning computing and build stronger university/college/K12 partnerships. Through this supplement, we will extend the Alliance with at least three new STARS Computing Corps, providing leadership training to a group of 8-10 students in each Corps, all focused on supporting the BJC effort. New Corps will provide teaching assistance to high school teachers implementing the BJC course through classroom visits and monthly Computer Science Teacher Association chapter meetings. These new STARS Computing Corps will also teach BJC material either through in middle school Citizen Schools after-school programs, and K-12 summer camps. This will provide a vibrant community of support for high school teachers and students engaging the new BJC course.

The Beauty and Joy of Computing project presents a unique opportunity to scale the STARS Alliance further while also enhancing national efforts to engage more high school teachers and students in teaching and learning computing and build stronger university/college/K12 partnerships. Through this supplement, we will extend the Alliance with at least three new STARS Computing Corps, providing leadership training to a group of 8-10 students in each Corps, all focused on supporting the BJC effort. New Corps will provide teaching assistance to high school teachers implementing the BJC course through classroom visits and monthly Computer Science Teacher Association chapter meetings. These new STARS Computing Corps will also teach BJC material either through in middle school Citizen Schools after-school programs, and K-12 summer camps. This will provide a vibrant community of support for high school teachers and students engaging the new BJC course.

Creating intelligent learning technologies from data has unique potential to transform the American educational system, by building a low cost way to adapt learning environments to individual students, while informing research on human learning. This project will create the technology for a new generation of data-driven intelligent tutors, enabling the rapid creation of individualized instruction to support learning in science, technology, engineering, and mathematics (STEM) fields. This has the potential to make individualized learning support accessible for a broad audience, from children to adults, including students that are traditionally underrepresented in STEM fields. This project will (1) develop computational methods to derive cognitive models from data that can be used to support individual learners through guidance, feedback, and help; (2) develop approaches to providing student support that leverage data to provide hints and guidance based on information such as frequency of student responses, probability of future errors, and solution efficiency; (3) develop interactive visualization tools for teachers to learn from student data in real time, to allow teachers and instructional designers to tailor instruction to address actual, rather than perceived, student problem areas; and (4) conduct formal empirical evaluations of the pedagogical effectiveness of our student support. Our software will construct adaptive support for teaching and learning in logic, discrete mathematics, and other STEM domains using a data-driven approach. From the extensive but tractable student performance data in computer-aided learning environments, we will automatically construct student cognitive models. Our cognitive models will build on our prior work using Markov Decision Processes and dimensionality reduction methods that leverage past data to assess student performance, direct a student’s learning path, and provide contextualized hints. We will use machine learning techniques to expand our problem-specific models into more general cognitive models to bootstrap the construction of new tutors and learn about student learning. For teachers and learning researchers, we will build a web-based visualization and analysis tool to graphically and interactively model student solutions annotated with performance data that reflects frequency, tendency to commit future errors, and closeness to a final solution. Through our new tutors and tools we will conduct experiments to understand student learning in a variety of contexts and domains, including logic, algebra, and chemistry. We will engage a team of diverse students and colleagues to bring interdisciplinary expertise to our research and share our findings broadly. This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

Creating a computationally literate citizenry is critical to maintaining national technological and scientific strength in a global society. While efforts have been underway to improve computing education, increase interest in computer science, and broaden participation in computing, these efforts have targeted the middle school and high school levels. Research suggests that there are concepts and practices in computational thinking that are readily graspable by elementary age students. This project will address this significant need by developing and iteratively refining a computer science curriculum to introduce computer science to elementary school students. The objectives of the project are 1) Extract best practices for elementary CS Education; 2) Iteratively define a multi-week computer science curriculum targeting 4th and 5th grade students. Iteratively define a multi-week computer science curriculum targeting 4th and 5th grade students; and 3) Pilot the curriculum, including professional development, within Conn Elementary School in Raleigh, North Carolina.

While great resources have been committed to developing computer science curricula at the university and high school levels, much less time has been devoted to developing curricula for younger students. The time is ripe to extend effective computer science pedagogy into the elementary realm. This project will develop and pilot a six-week elementary computer science curriculum focusing on problem solving, creativity, and computer science principles.

In order to maintain its position in a global society, the US must strengthen and diversify its computer science workforce. For training computer scientists, collaborative learning has become an increasingly important approach. Despite substantial evidence of the overall benefits of collaborative learning and pair programming in particular, it is not yet known how the fine-grained facets of natural language dialogue contribute to the effectiveness of collaboration. Additionally, there are many open questions regarding the ways in which collaborative dialogue can foster a sense of computing identity, motivation, and engagement for students from underrepresented groups while supporting rigorous computer science learning. The proposed project will study collaborative learning in the second computing course for majors in order to build fine-grained computational models of their collaboration. Predictive models will be built to explain the extent to which particular dialogue phenomena contribute to learning and motivational outcomes. Finally, pedagogical interventions will be proposed and investigated based upon those computational models. The results are expected to serve as the basis of a grounded theory of how collaborative dialogue for computer science education can support rigorous learning and diverse student retention.

Internal award supplement

Human one-on-one tutoring is one of the most effective educational interventions. Tutored students often perform significantly better than students in classroom settings (Bloom, 1984; Cohen, Kulik, & Kulik, 1982). Computer learning environments that mimic aspects of human tutors have also been highly successful. Intelligent Tutoring Systems (ITSs) have been shown to be highly effective in improving students' learning at real classrooms (Anderson, Corbett, Koedinger, & Pelletier, 1995; Koedinger, Anderson, Hadley, & Mark, 1997; VanLehn et al., 2005). The development of ITSs has enabled schools and universities to reach out and educate students who otherwise would be unable to take advantage of one-on-one tutoring due to cost and time constraints (Koedinger, Anderson, Hadley, & Mark, 1997). Despite the high payoffs provided by ITSs, significant barriers remain. High development costs and the challenges of knowledge engineering have prevented their widespread deployment. A diverse team of software developers, domain experts, and educational theorists are required for development, testing and even maintenance. Generally speaking, it requires an average of 80 man-hours per hour of tutoring content. In this proposed work, our goal is to automatically design effective personalized ITSs directly from log data. We will combine co-pI Dr. Barnes data-driven approach on learning what to teach with PI Dr. Chi’s data-driven work on learning how to teach. More specifically, we will explore two important undergraduate stem domains: discrete math and probability; and employ two types of ITSs: an example-based ITS, the discrete math tutor, and a rule-based ITS, Pyrenees. The former can automatically generate hints directly from students’ prior solutions while the latter has hard-coded domain rules and teaches students a domain-general problem-solving strategy within the context of probability. For learning how to teach, we will apply reinforcement learning to induce adaptive pedagogical strategies directly from students’ log files and will focus on three levels of pedagogical decisions: 1) whether or not to give students hints and what level of hint (step-level); 2) whether to show students worked example or ask students to engage problem solving (problem-level); and 3) whether or not to teach students a meta-cognitive problem-solving strategy (metacognitive level).

Every year approximately 700,000 people die in US hospitals. In 16% of them, the first diagnosis at death was septicemia – one of the most common delayed diagnoses associated with inpatient death. Sepsis is one of the ten leading causes of death. While it is difficult to estimate how many of these deaths could have been averted or postponed if a better system of care was in place, it is widely recognized that sepsis patients have better outcomes if treated earlier. Sepsis is one of the most common of these diagnoses with delayed effective treatment interventions. As opposed to wrong diagnoses, delayed diagnoses have historically not been considered adverse events as there is no change in patient condition as a result of care delivered. However, patients with delayed diagnoses do have worse outcomes than those who receive timely treatment. These diagnostic and/or treatment delays associated with inpatient mortality and long term morbidity consequences represent a significant and modifiable patient safety issue. Awareness of sepsis is low; many septic patients are under-diagnosed at an early stage when aggressive treatment could still reverse the course of the infection. Early recognition and implementation of early goal directed therapy improves outcomes and decreases mortality. For every one hour delay in treatment of severe sepsis or severe shock with antibiotics, there is an incremental decrease in patient survival. For example, a delay in antibiotics of five hours decreases survival to 50%. We propose to take a date-driven evidence-based approach that integrates computer science and industrial engineering to develop personalized sepsis diagnosis and treatment plans. The goal of this research is to integrate electronic health records (EHR) and clinical expertise to provide an evidence-based framework for diagnosing sepsis patients, accurately risk-stratifying patients within the sepsis spectrum, and developing intervention policies that inform sepsis treatment decisions. We will achieve this research goal through accomplishing three specific aims based on a longitudinal data set of EHRs of Mayo Clinic Rochester hospital patients discharged and Christiana Care Health System hospital patients.

The objective of this letter of intent is to indicate that North Carolina State University (NCSU) will join, as a site, the Center of Hybrid Multicore Productivity Research (CHMPR) in Year 2 of its planned phase II I/UCRC renewal. CHMPR consists of sites at the University of Maryland, Baltimore County, and the University of California, San Diego. CHMPR is proposing a phase II extension of its successful conduct of research based on the continued challenges arising from the evolution of multicore technologies. The focus of NCSU within the center will be the science of technologies for end-to-end enablement of data. We are forming an NCSU-based site to join an existing I/UCRC Center for Hybrid Multicore Productivity (CHMPR). This planning grant will support meetings with potential industrial partners and CHMPR co-investigators from the University of Maryland Baltimore County. Planning activities will be facilitated by the NCSU CSC Director of Development & External Relations, Corporate & Alumni Relations Mr. Ken Tate, as well as by the Director of Institute for NEXT generation IT Systems Mr. John Streck. Meetings organized using the planning grant funds will operationalize, transform and integrate themes developed during the preparation of this planning grant, experiences from ongoing industry partnerships at the three centers, and concepts outlined in NSF’s ‘Purple Book’ entitled Managing the Industry/University Cooperative Research Center. The NCSU I/UCRC center site will represent an exciting and novel integration of approaches to Big Data in multiple disciplines, as we target technological innovations on extracting value from Big Data for a diverse pool of stakeholders.

DO3 Task 2.7 Data Readiness

Software engineers need improved tools and methods for translating complex, changing legal regulations into workable information technology systems. Compliance with legal requirements is an essential element in trustworthy systems. This project will apply formal methods to advance the cutting edge for creating more accurate, efficient, and reliable RCSE, resulting in compliant software systems.

DO3 Task Order 2.9 KRM

The explosive growth in bandwidth represented by advances in optical communication and networking technologies has underpinned the increasing reach and reliability of the Internet in the last two decades. However, the potential impact of increasingly sophisticated recent advances in optical technology, such as rapid switching and elastic wavelengths have not yet been realized. The main cause of this is that such technology, while possible to integrate into the data plane of planetary networking, is difficult to accommodate in the current planning, management, and control strategies. We propose in this project to work hand-in-hand with collaborating researchers from NICT, Japan, who are working to realize a novel technology of hybrid optical packet/circuit switching. Such a technology could be immensely useful to large transport network operators, but there are no existing algorithms that can easily determine how a provider can provision their resources between the circuit and packet possibilities on an ongoing dynamic basis. We envision a novel approach to this problem, where we utilize the concept of a "choice marketplace" that allows sophisticated rendezvous semantics between customer and provider, and allows them to cooperatively guide network resource provisioning to dynamically fulfill network objectives such as maximizing performance received by network traffic. Our approach also allows balancing of various objectives, such as network utilization, latency, and the increasingly important metric of energy expenditure in the network.

Cognitive radio is an emerging advanced radio technology in wireless access, with many promising benefits including dynamic spectrum sharing, robust cross-layer adaptation, and collaborative networking. Opportunistic spectrum access (OSA) is at the core of cognitive radio technologies, which has received great attention recently, focusing on improving spectrum utilization efficiency and reliability. However, the state-of-the-art still suffers from one severe security vulnerability, which has been largely overlooked by the research community so far. That is, a malicious jammer can always disrupt the legitimate network communication by leveraging the public-available channel statistic information to effectively jam the channels and thus lead to serious spectrum underutilization. In this proposal, we propose to address the challenge of effective anti-jamming communication in cognitive radio networks (CRNs). We propose a multiple lines of defense approach, which considers and integrates defense technologies from different dimensions, including frequency hopping, power control, cooperative communication, and signal processing. The proposed defense approach enables both reactive and proactive protection, from evading jammers to competing against jammers, and to expelling jamming signals, and thus guarantees effective anti-jamming communication under a variety of network environments.

This project builds on the SILO project that started in 2006 to design a new architecture for the Internet. In this new project, we will collaborate with teams of researchers from the University of Kentucky, the University of Massachusetts, and RENCI, to design critical parts of a new architecture for the Internet that will support the flexible use of widely applicable information transport and transformation modules to create good solutions for specific communication applications. The key idea is to allow a network to offer information transformation services at the edge or in the core transparently to the application, and creating a framework in which application can issue a request not only for communication but for specific reusable services. We also propose research tasks that will enable network virtualization and isolation seamlessly at many levels, currently a difficult but highly relevant problem in practical networking.

Smartphones and mobile devices have become a dominant computing platform for billions of users. There are many static and dynamic program analysis tools to detecting security and privacy concerns in mobile applications. However, few approaches bridge the semantic gap between code and visual presentation. Ignoring this context results in analysis that falsely reports an application as malicious (e.g., the user really wanted to use an app to record phone calls), or fails to detect suspicious behavior (e.g., an app that collects sensitive information via text input). We propose to use a hybrid static / dynamic approach to extract the text labels from the Android application UIs followed by text analytics to semantically label the type of input asked for by the application. Doing so will better characterize the amount of security and privacy information entered into Android applications, as well as enable outlier detection to identify applications that ask for unexpected (e.g., SSN) information for their semantic class (e.g., banking applications). This analysis will be applied at scale to identify potential privacy infringements in mobile application stores.

Computing systems that make security decisions often fail to take into account human expectations. This failure occurs because human expectations are commonly drawn from textual sources (e.g., mobile application description) and are hard to extract and codify. This proposal seeks to extract expectation context from natural-language text artifacts presented to users as they find, install, and run software. The proposed work focuses specifically mobile applications to demonstrate concrete and practical advances in our scientific understanding of applying user expectation context to security decisions. These findings will advance the state-of-the-art in identifying and classifying malware and grayware, as well as identify better methods of communicating risk to users. We will also gain a better understanding of the unique challenges of applying text analytics the security domain.

Controlling the access and use of information is a fundamental challenge of computer security. Emerging computing platforms such as Android and Windows 8 further complicate access control by relying on sharing and collaboration between applications. When more than two applications participate in a workflow, existing permission systems break down due to their boolean nature. In this proposal, we seek to provide applications with residual control of their data and its copies. To do this, we propose secure OS views, which combines a new abstraction for accessing data with whole-system information tracking. We apply secure OS views to modern operating systems (e.g., Android and Windows 8), which use database-like abstractions for sharing and accessing information. Similar to a database view, secure OS views uses runtime context to dynamically define the protection domain, allowing the return of the value, a fake value, or nonexistence of the record.

Mobile browsers are beginning to serve as critical enablers of modern computing. With a combination of rich features that rival their desktop counterparts and strong security mechanisms such as TLS/SSL, these applications are becoming the basis of many other mobile apps. Unfortunately, the security guarantees provided by mobile browsers and the risks associated with today's mobile web have not been evaluated in great detail. In the proposed work, we will investigate the security of mobile browsers and the mobile specific websites they access. Characterizing and responding to the threats in this space is critical, especially given that cellular devices are used by more than five billion people around the world

Title: Refining Security Policy for Smartphone Applications Abstract: Smartphones running Android and iOS have penetrated all areas of the public and private sectors. These platforms mark an important milestone in operating system security: they enforce security policies on applications, not users. Both platforms use least privilege security goals for user-applications downloaded from application markets. However, whereas Android largely enforces security goals with OS policy, iOS depends highly on a review process. Unfortunately, review alone can be circumvented, as demonstrated by multiple recent works. This proposal seeks to deepen our understanding of least privilege policy for user-applications by designing formal frameworks and algorithms for modeling, extracting, and enforcing policy. Successful execution of the proposed tasks enhance end-user security and privacy in Android and iOS devices.

Increased use of cloud computing services is becoming a reality in today's IT management. The security risks of this move are active research topics, yielding cautionary examples of attacks enabled by the co-location of competing tenants. In this project, we propose to mitigate such risks through a new approach to cloud architecture defined by leveraging cloud providers as trusted (but auditable) security enablers. We will exploit cooperation between cloud providers and tenants in preventing attacks as a means to tackle long-standing open security problems, including protection of tenants against outsider attacks, improved intrusion detection and security diagnosis, and security-monitoring inlays.

Peer review between students has a 40-year history in academia. During the last half of that period, web-based peer-review systems have been used in tens of thousands of classes. Many online systems have been developed, in diverse settings and with diverse purposes. The systems, however, have common concerns: assigning capable reviewers to each student submission, insuring review quality, and delivering reliable scores, in cases where the systems are used for summative review of student work. Many strategies have been proposed to meet those concerns, and tested in relatively small numbers of courses. The next step is to scale up the studies to learn how well they perform in diverse settings, and with large numbers of students. This project brings together researchers from several peer-review systems, including some of the largest, to build web services that can be incorporated into existing systems to test these strategies and visualize the results.

The students assist our efforts to build a database of peer-review responses that can be mined for quantitative research studies. The database will be composed of anonymized data from the peer-review systems of the constituent projects: CritViz, CrowdGrader, Expertiza, and Mobius/Slip. Among other items, it will contain peer feedback and ratings, and links to submitted work. They will embark on a qualitative research study to determine what STEM students value about the peer-review process. They will use a common set of research protocols to investigate three research questions: What do students value about receiving reviews? What do they value about giving reviews? Do their reactions differ, based on demographics, age/level of study, or academic major?

Hosting infrastructures provide users with cost-effective computing solutions by obviating the need for users to maintain complex computing infrastructures themselves. Unfortunately, due to their inherent complexity and sharing nature, hosting infrastructures are prone to various system anomalies caused by various external or internal faults.The goal of this project is to investigate a holistic,cross-site, hybrid system anomaly debugging framework that intelligently integrates production-site black-box diagnosis and developer-site white-box analysis into a more powerful hosting infrastructure anomaly debugging system.

Large-scale virtualized hosting infrastructures have become the fundamental platform for many real world systems such as cloud computing, enterprise data centers, and educational computing lab. However, due to their inherent complexity and sharing nature, hosting infrastructures are prone to various runtime problems such as performance anomalies and software/hardware failures. The overarching objective of this proposal is to systematically explore innovative runtime reliability management techniques for large-scale virtualized hosting infrastructures. Our research focuses on handling performance anomalies in distributed systems that are often very difficult to reproduce offline. Our approach combines the power of online learning, knowledge-driven first-response recovery, and in-situ diagnosis to handle unexpected system anomalies more efficiently and effectively. We aim at transforming the runtime system anomaly management from a trial-and-error guessing game into an efficient knowledge-driven self-healing process.

Cloud computing infrastructure provides an elastic application deployment environment. Applications can be dynamically instantiated on different physical hosts on demand. However, in order to fully explore the elasticity of the cloud infrastructure, applications should be able to automatically configure themselves when their components are placed in or migrated to different data centers at geographically distributed regions. Unfortunately, today’s technology does not provide such an automatic configuration support. The application developer still needs to deal with the hassle of configuring their applications manually. The objective of this proposed project is to develop an automatic application configuration management framework that can decouple the configuration management from the application logic. I propose to conduct the following tasks: 1) identifying the configuration problems in existing cloud platform and container model; 2) Designing the configuration management framework and interfaces that can decouple the configuration management from the application. We will use a multi-tier online auction application as a case study example; and 3) developing our designed configuration management framework and optimizing its performance.

Our goal in this project is to focus on two related tasks. First, we will extend work from last year’s (2014-2015) SAS-NCSU MRA research project to study predictive analytics for narrative threads: the ability to predict how a sequence of narrative events could unfold in the future, based on past and current events; algorithms to determine how different past choices would have effect on current and future states; and visualization and user interface tools and techniques to allow users to explore the predictive space in efficient and effective ways. Second, we will collaborate with the deep learning group to study the problem of visualizing nodes and relationships in the layers of a convolution neural networks (CNN). Visualization may allow researchers to gain important insights into how different nodes in the layers of a CNN are contributing to their final output.

This project will investigate the use of text analytics and information visualization for analyzing, visualizing, and exploring large document collections. Text will be analyzed using a combination of SAS Text Miner and new algorithms designed to identify, structure, and associate context, sentiment, and other properties in a document collection. Results will be segmented into narrative threads that identify conceptual storylines within the document collection. Threads will be visualized using two approaches: node-link graphs and adjacency matrices. For the graphs, graph analysis algorithms will be studied to determine methods to derive useful metrics and summaries on the document collection. For the adjacency matrices, algorithms will be developed to order of rows and columns in ways that expose patters in the document collection. For both visualization techniques, issues of multivariate data representation and level-of-detail hierarchies will also be considered.

This proposal is for the development and evaluation of CAPTIVE, a Cube with Augmented Physical Tools, to support exploration of three-dimensional information. The design of CAPTIVE is founded on the concept of tool use, in which physical objects (tools) are used to modify the properties or presentation of target objects. CAPTIVE integrates findings across a wide range of areas in human-computer interaction and visualization, from bimanual and tangible user interfaces to augmented reality. CAPTIVE is configured as a desktop augmented reality/fishtank virtual reality system [120], with a stereo- scopic display, a haptic pointing device, and a user-facing camera. In one hand the user holds a wireframe cube that contains virtual objects, in the other the pointing device, augmented to reflect its function as a tool: a probe probes for pointing at, choosing, and moving objects; a magnifying or semantic lens for filter- ing, recoding, and elaborating information; a cutting plane that shows slices or projection views. CAPTIVE supports visualization with more fluid and natural interaction techniques, improving the ability of users to explore and understand 3D information.

Plants, as sessile organisms, need to constantly adjust their intrinsic growth and developmental programs to the environmental conditions. These environmentally triggered “adjustments“ often involve changes in the developmentally predefined patterns of one or more hormone activities. In turn, these hormonal changes result in alterations at the gene expression level and the concurrent alterations of the cellular activities. In general, these hormone-mediated regulatory functions are achieved, at least in part, by modulating the transcriptional activity of hundreds of genes. The study of these transcriptional regulatory networks not only provides a conceptual framework to understand the fundamental biology behind these hormone-mediated processes, but also the molecular tools needed to accelerate the progress of modern agriculture. Although often overlooked, understanding of the translational regulatory networks behind complex biological processes has the potential to empower similar advances in both basic and applied plant biology arenas. By taking advantage of the recently developed ribosome footprinting technology, genome-wide changes in translation activity in response to ethylene were quantified at codon resolution, and new translational regulatory elements have been identified in Arabidopsis. Importantly, the detailed characterization of one of the regulatory elements identified indicates that this regulation is NOT miRNA dependent, and that the identified regulatory element is also responsive to the plant hormone auxin, suggesting a role in the interaction between these two plant hormones. These findings not only confirm the basic biological importance of translational regulation and its potential as a signal integration mechanism, but also open new avenues to identifying, characterizing and utilizing additional regulatory modules in plants species of economic importance. Towards that general goal, a plant-optimized ribosome footprinting methodology will be deployed to examine the translation landscape of two plant species, tomato and Arabidopsis, in response to two plant hormones, ethylene and auxin. A time-course experiment will be performed to maximize the detection sensitivity (strong vs. weak) and diversity (early vs. late activation) of additional translational regulatory elements. The large amount and dynamic nature of the generated data will be also utilized to generate hierarchical transcriptional and translational interaction networks between these two hormones and to explore the possible use of these types of diverse information to identify key regulatory nodes. Finally, the comparison between two plant species will provide critical information on the conservation of the regulatory elements identified and, thus, inform research on future practical applications. Intellectual merit: The identification and characterization of signal integration hubs and cis-regulatory elements of translation will allow not only to better understand how information from different origins (environment and developmental programs) are integrated, but also to devise new strategies to control this flow for the advance of agriculture. Broader Impacts: A new outreach program to promote interest among middle and high school kids in combining biology, computers, and engineering. We will use our current NSF-supported Plants4kids platform (ref) with a web-based bilingual divulgation tools, monthly demos at the science museum and local schools to implement this new outreach program. Examples of demonstration modules will include comparison between simple electronic and genetic circuits.

The coordination of spatial patterning cues and cellular proliferation underlies diverse processes from cancerous growth to reproductive development. A long-term objective of my research program is to understand how proliferative cues are coordinated with spatial information during organogenesis. In Arabidopsis thaliana this coordination of patterning and proliferation is necessary within the carpel margin meristem (CMM) to generate ovules that when fertilized will become seeds. In the previous funding period we demonstrated that the SEUSS (SEU) and AINTEGUMENTA (ANT) transcription factors regulate critical patterning events that support carpel margin meristem and ovule development. Our genetic analysis demonstrates that SEU and ANT share a partially redundant and overlapping function essential for proper seed formation. As SEU and ANT do not share sequence similarity, the molecular basis for this redundancy is not understood. We propose that the SEU and ANT activities synergistically converge at key transcriptional nodes. A node in this sense is a gene or a set of related genes that requires the combined activities of SEU and ANT for its expression. Our recently published transcriptomic analysis indicates that many of these nodes encode known transcriptional regulators. By studying these nodes we hope to better understand the transcriptional hierarchies that control CMM development and uncover the mechanistic basis of the synergistic action of SEU and ANT. Our transcriptomics study cannot determine if the nodes that we have identified are directly or indirectly regulated by SEU or ANT activity, However, even if these genes are indirectly controlled by SEU and ANT activity, their expression within the developing CMM suggests they may still play a critical functional role during CMM development. Furthermore, having now identified a set of genes that are enriched for CMM expression we are in a position to study the cis-regulatory elements that support gene expression within the CMM and medial gynoecial domain. Thus here we propose to: 1) Identify direct targets of SEU regulation within the CMM to further refine the transcriptional hierarchy required for CMM development; 2) assay the functional role of two of these nodes during CMM development; one encoded by the transcription factor PERIANTHIA and the second encoded by members of the REM family of B3 domain-containing proteins; 3) Identify cis-acting DNA regulatory elements required for CMM expression. Scientific significance: Understanding the coordination of cellular proliferation and spatial patterning during organogenesis is broadly of interest to scientists working in a diversity of fields. Completion of these specific aims will move us toward this future goal by illuminating the mechanistic basis for the overlapping functions of SEU and ANT during carpel margin and ovule development. Additionally, we expect that by elucidating the molecular mechanisms of the synergistic action of SEU and ANT upon key transcriptional nodes, we will engender a greater understanding of the molecular underpinnings of non-additivity within transcriptional networks and the complexity of developmental programs. Past NSF funding for this project (IOS-0821896) has resulted in the publication of five articles in well-respected journals (two in Plant Physiology, and one each in Developmental Biology, PLoS One, and BMC Plant Biology). Broader impacts: I ensure a broad societal impact from my program by integrating my research efforts with my teaching and training responsibilities and by widely disseminating materials and results. Furthermore, I initiated and continue to lead an outreach group that prepares and presents hands-on science demonstrations at local North Carolina schools. Our group has reached over 1500 Kindergarten through Grade 12 students over the past six years and continues to develop new demonstration modules inspired by our current work in developmental biology and genetics.

The computer science education (CSEd) research community consists of a large group of passionate CS educators who often contribute to other disciplines of CS research. There has been a trend in other disciplines toward more rigorous and empirical evaluation of various hypotheses. However, many of the practices that we apply to demonstrate rigor in our discipline research are ignored or actively avoided when performing research in CSEd. This suggests that CSEd is “theory scarce” because most publications are not research and do not provide the evidence or replication required for meta-analysis and theory building . An increase in empiricism in CSEd research will move the field from “scholarly teaching” to the “scholarship of teaching and learning” (SoTL) providing the foundation for meta-analysis and the generation of theories about teaching and learning in computer science. We propose the creation of training workshops and tutorials to educate the educators about appropriate research design and evaluation of educational interventions. The creation of laboratory packages, “research-in-a-box,” will support sound evaluation and replication leading to meta-analysis and theory building in the CSEd community.

Students are first exposed to the full software engineering lifecycle in CSC216, a second semester programming course taught in Java. To support engagement and retention of our students, we hypothesize that teaching best software engineering practices and modeling successful techniques for solving computer science and software engineering problems in the classroom will support student success on independent programming projects for CSC216 and future courses. The software engineering best practices of interest are: test-driven development, coverage, static analysis, inspections, version control, and continuous integration. This project includes the development, integration, and refinement of instructional materials for engaging students in software engineering best practices during in-class laboratory assignments. We will assess the effectiveness of the instructional materials and delivery by measuring student learning, engagement, efficacy, and retention.

Stealthy rootkit attacks are one of the most foundational threats to cyberspace. With the capability of subverting the software root of trust of a computer system, i.e., the operating system (OS) or the hypervisor, a rootkit can instantly take over the control of the system and stealthily inhabit the victim. To effectively mitigate and defeat them, researchers have explored various solutions. Unfortunately, the state-of-the-art defense is mainly reactive and in a fundamentally disadvantageous position in the arms-race against these stealthy attacks. The proposed research aims to fundamentally change the arms-race by proposing a systematic immunization approach to proactively prevent and exterminate rootkit attacks. Inspired by our human immune system and fundamental biological design principles, the proposed approach transforms system software (i.e., the OS and the hypervisor) so that the new one will tip the balance of favor toward the rootkit defense. To accomplish that, we will investigate a suite of innovative techniques to preserve kernel/hypervisor control flow integrity and evaluate their effectiveness with real-world malware and infrastructures. The proposed education components include the creation and dissemination of unique hands-on course materials with live demos, lab sessions, and tutorials.

Contemporary research on multi-agent learning environments has focused on self-regulated learning (SRL) while relatively little effort has been made to use co-regulated learning as a guiding theoretical framework (Hadwin et al., 2011). This oversight needs to be addressed given the complex nature that self-and other-regulatory processes play when human learners and artificial pedagogical agents (APAs) interact to support learners? internalization of cognitive, affective, and metacognitive (CAM) SRL processes. We will use the Crystal Island learning environment to investigate these issues.

Although the majority of adolescent health problems are amenable to behavioral intervention, and most adolescents are comfortable using interactive computing technology, few health information technology interventions have been integrated into adolescent care. The objective of the proposed research is to design, implement, and investigate INSPIRE, a self-adaptive personalized behavior change system for adolescent preventive healthcare. With a focus on adolescents, INSPIRE will enable adolescents to be active participants in dynamically generated, personalized narrative experiences that operationalize theoretically grounded interventions for behavior change through interactive narratives? plot structures and virtual character interactions.

Despite the great promise offered by learning environments for K-12 science education, realizing its potential poses significant challenges. In particular, current learning environments do not adaptively support children's affect. This project investigates computational models of affect that support multimodal dialogue-rich interaction. With an emphasis on devising scalable solutions, the project focus on machine-learning techniques for automatically acquiring affect and dialogue models that can be widely deployed in natural classroom settings.

The goal of the ENGAGE project is to develop a game-based learning environment that will support middle grade computer fluency education. It will be conducted by an interdisciplinary research team drawn from computer science, computer science education, and education. In collaboration with North Carolina middle schools, the research team will design, develop, deploy, and evaluate a game-based learning environment that enables middle school students to acquire computer fluency knowledge and skills. The ENGAGE project will be evaluated in middle grade classrooms with respect to both learning effectiveness and engagement.

The project will develop automated detectors that can infer the engagement and affect of trainees learning through the vMedic training system. This project will combine interaction-based methods for detecting these constructs (e.g., models making inference solely from the trainee?s performance history) with scalable sensor-based methods for detecting these constructs, towards developing models that can leverage sensor information when it is available, but which can still assess trainee engagement and affect effectively even when sensors are not available. The automated detectors will be developed, integrated together, and validated for accuracy when applied to new trainees.

An important problem in Natural Language Processing is creating a robust language model that represents how a set of documents is composed from a set of words. Having a robust language model is critical for many text analytic tasks, including topic modeling, information retrieval, text categorization, and sentiment analysis. Most language models utilize a bag-of-words representation that ignores word order. While these representations suffice for some text analytic tasks, a context-sensitive language model is essential for many tasks such as discovering semantic relationships, question answering, machine translation, and dialog modeling. This project will investigate the effectiveness of leveraging Deep Learning architectures to create context-sensitive language models.

The goal of the Leonardo project is to develop an intelligent cyberlearning system for interactive scientific modeling. Students will use Leonardo's intelligent virtual science notebooks to create and experiment with interactive models of physical phenomena. As students design and test their models, Leonardo's intelligent virtual tutors will engage them in problem-solving exchanges in which they will interactively annotate their models as they devise explanations and make predictions. During the project, the Leonardo virtual science notebook system will be rolled out to 60 classrooms in North Carolina, Texas, and California.

Intelligent tutoring systems (ITSs) are highly effective for education and training. Tutorial planning is a critical component of ITSs, controlling how scaffolding is structured and delivered to learners. Devising data-driven tutorial planners that automatically induce scaffolding models from corpora of student data holds significant promise. This project investigates a data-driven framework for tutorial planning that is based on modular reinforcement learning. This framework explicitly accounts for the inherent uncertainty in how learners respond to different types of tutorial strategies and tactics, and automatically induces and refines tutorial planning policies in order to optimize measures of learning outcomes.

Intelligent virtual humans (IVHs) are able to connect with real people in powerful, meaningful, and complex ways. They can mimic the behavior of real people and therefore add a rich social dimension to computer interactions, providing not only a wealth of information but presenting information in more personals ways. This 3-year project will focus on testing the effectiveness of IVHs in facilitating college students self-regulated learning in STEM with MetaTutor. More specifically, we plan to test IVHs detection, monitoring, and modeling (both facially and verbally) the temporal dynamics of learners self-regulatory processes to enhance learners deployment of effective learning strategies, accurate metacognitive judgments, and appraisals of emotional states. This will be accomplished by aligning and conducting complex computational and statistical analyses of a multitude of trace data (e.g., log-files, eye-tracking), behavioral (e.g., human-virtual human dialogue moves), physiological measures (e.g., GSR, ECG, EEG), and learning outcome data collected in real-time. The proposed research, in the context of using IVHs, is extremely challenging and will help us to better understand the nature and temporal dynamics of these processes, how they contribute to various types of learning, and provide the empirical basis for designing intelligent virtual human systems. The results from this grant will contribute significantly to models and theories of social, cognitive, and physiological bases of human-virtual human interactions; statistical and computational methods to make inferences from complex multi-channel data; theoretical and conceptual understanding of temporally-aligned data streams, and enhancing students understanding of complex science topics by making more sensitive and intelligent virtual humans.

This project will utilize research and development groups at the Concord Consortium and NC State University. Educational software for teaching high school multi-level genetics developed by the Concord Consortium will be enhanced by intelligent agents and machine-based tutoring system technologies developed at NC State to help enhance the learning experience for students. These groups will collaborate closely to develop and research a hybrid system that combines technological intervention and teacher pedagogical expertise to illuminate and guide student learning in deeply digital curricula and classrooms.

Software engineers need better ways to recognize best practices in past projects, and to understand how to transfer and adapt those experiences to current projects. No project is exactly like previous projects- hence, the trick is to find which parts of the past are most relevant and can be transferred into the current project. We propose novel automated methods to apply the machine learning concept of transfer learning to adapt lessons from past software engineering project data to new conditions.

Software cost estimation remains a on-going open issue at NASA. It is notoriously difficult to estimate the development costs of large complex deep space science experiments. To support NASA's Jet Propulsion Laboratory, NcState is applying state of the art spectral clustering methods to improve case-based reasoning.

Multicore platforms have the potential of revolutionizing the capabilities of embedded cyber-physical systems but lack predictability in execution time due to shared resources. Safety-critical systems require such predictability for certification. This research aims at resolving this multicore ``predictability problem.'' It will develop methods that enable to share hardware resources to be allocated and provide predictability, including support for real-time operating systems, middleware, and associated analysis tools. The devised methods will be evaluated through experimental research involving synthetic micro-benchmarks and code for unmanned air vehicles ``re-thinking'' their adapting to changing environmental conditions within Cyber-Physical Systems.

This project intends to deliver an operating system and runtime system (OS/R) environment for extreme-scale scientific computing. We will develop the necessary OS/R interfaces and lowlevel system services to support isolation and sharing functionality for designing and implementing applications as well as performance and correctness tools. We propose a lightweight OS/R system with the flexibility to custom build runtimes for any particular purpose. Each component executes in its own "enclave" with a specialized runtime and isolation properties. A global runtime system provides the software required to compose applications out of a collection of enclaves, join them through secure and low-latency communication, and schedule them to avoid contention and maximize resource utilization. The primary deliverable of this project is a full OS/R stack based on the Kitten operating system and Palacios virtual machine monitor that can be delivered to vendors for further enhancement and optimization.

This decade is projected to usher in the period of exascale computing with the advent of systems with more than 500 million concurrent tasks. Harnessing such hardware with coordinated computing in software poses significant challenges. Production codes tend to face scalability problems, but current performance analysis tools seldom operate effectively beyond 10,000 cores. We propose to combine trace analysis and in-situ data analysis techniques at runtime. Application developers thus create ultra low-overhead measurement and analysis facilities on-the-fly, customized for the performance problems of particular application. We propose an analysis generator called ScalaJack for this purpose. Results of this work will be contributed as open-source code to the research community and beyond as done in past projects. Pluggable, customization analysis not only allows other groups to build tools on top of our approach but to also contribute components to our framework that will be shared in a repository hosted by us.

In High-End Computing (HEC), faults have become the norm rather than the exception for parallel computation on clusters with 10s/100s of thousands of cores. As the core count increases, so does the overhead for fault-tolerant techniques that rely on checkpoint/restart (C/R) mechanisms. At 50% overheads, redundancy is a viable alternative to fault recovery and actually scales, which makes the approach attractive for HEC. The objective of this work to the develop a synergistic approach by combining C/R-based fault tolerance with redundancy in computer to achieve high levels of resilience. This work alleviates scalability limitations of current fault tolerant practices. It contributes to fault modeling as well as fault detection and recovery in significantly advancing existing techniques by controlling levels of redundancy and checkpointing intervals in the presence of faults. It is transformative in providing a model where users select a target failure probability at the price of using additional resources.

he objective of this work is to provide functionality for the BLCR Linux module under a PGAS runtime system (within the DEGAS software stack) to support advanced fault-tolerant capabilities, which are of specific value in the context of large-scale computational science codes running on high-end clusters and, ultimately, exascale facilities. Our proposal is to develop and integrate into DEGAS a set of advanced techniques to reduce the checkpoint/restart (C/R)overhead.

The objective of this work is to provide functionality for the BLCR Linux module under a PGAS runtime system (within the DEGAS software stack) to support advanced fault-tolerant capabilities, which are of specific value in the context of large-scale computational science codes running on high-end clusters and, ultimately, exascale facilities. Our proposal is to develop and integrate into DEGAS a set of advanced techniques to reduce the checkpoint/restart (C/R) overhead.

This proposal provides subcontractor support to Virginia Tech for a proposal submitted under the Air Force's Basic Research Initiative. The proposal will focus on development of reconfigurable mapping strategies for porting multi-block structured and unstructured-mesh CFD codes to computing clusters containing CPU/GPU processing units.

The objective of this work is to design and evaluate novel system and program abstractions for combined performance and scalability paving the path into a future of operating system supporting a massive number of cores on a single chip.

The objective of this NSF-CPS Synergy proposal is to develop a distributed algorithmic frame- work, supported by a highly fault-tolerant software system, for executing critical transmission-level operations of the North American power grid using gigantic volumes of Synchrophasor data. As the number of Phasor Measurement Units (PMU) increases to more than thousands in the next 4-5 years, it is rather intuitive that the current state-of-the-art centralized communication and information processing architecture of Wide-Area Measurement System (WAMS) will no longer be sustainable under such data-explosion, and a completely distributed cyber-physical architecture will need to be developed. The North American Synchrophasor Initiative (NASPI) is currently address- ing this architectural aspect by developing new communication and computing protocols through NASPInet and Phasor Gateway. However, almost no attention has yet been paid to perhaps the most critical consequence of this envisioned distributed architecture - namely distributed algorithms. Our primary task, therefore, will be to develop parallel computational methods for solving real-time wide-area monitoring and control problems with analytical investigation of their stability, convergence and robustness properties, followed by their implementation and testing against extraneous malicious attacks using our WAMS-RTDS testbed at NC State.

Software vulnerabilities originating from insecure code are one of the leading causes of security problems people face today. Current tool support for secure programming focuses on catching security errors after the program is written. We propose a new approach, interactive static analysis, to improve upon static analysis techniques by introducing a new mixed-initiative paradigm for interacting with developers to aid in the detection and prevention of security vulnerabilities.

Tools can help software developers alleviate the challenge of creating and maintaining software. Unfortunately, developers only use a small subset of the available tools. The proposed research investigates how social learning, an effective mechanism for discovering new tools, can help software developers to discover relevant tools. In doing so, developers will be able to increase software quality while decreasing development time.

LAS DO5 Murphy-Hill Task 5.4 Informatics

DO3 Task Order 2.11 Mission Enabling

DO3 Task Order 2.8 Analytic Workflow

A wide variety of program analysis tools have been created to help software developers do their jobs, yet the output of these tools are often difficult to understand and vary significantly from tool to tool. As a result, software developers may waste time trying to interpret the output of these tools, instead of making their software more capable and reliable. This proposal suggests a broad investigation of several types of program analysis tools, with the end goal being an improved understanding of how program analysis tools can inform developers in the most expressive and uniform way possible. Once this goal is reached, we can create program analysis tools that enable developers to make tremendous strides towards more correct, more reliable, and more on-time software systems.

This is a pilot internship program between NCSU and Cisco for 4 undergraduate students to learn through working part-time on real life problems for Cisco with the hope that this pilot program can grow and develop into a long term working relationship. Specifically, NCSU students will participate in Cisco Software Application Support plus Upgrades (SASU) projects and/or conduct research for SASU. This will be done with an understanding that the interns are students, and as such are learning and being trained with the training coming from both the Cisco (for SASU-specific skills), and NCSU (through the undergraduate program they are enrolled in) in general relevant skills.

Wireless security is receiving increasing attention as wireless systems become a key component in our daily life as well as critical cyber-physical systems. Recent progress in this area exploits physical layer characteristics to offer enhanced and sometimes the only available security mechanisms. The success of such security mechanisms depends crucially on the correct modeling of underlying wireless propagation. It is widely accepted that wireless channels decorrelate fast over space, and half a wavelength is the key distance metric used in existing wireless physical layer security mechanisms for security assurance. We believe that this channel correlation model is incorrect in general: it leads to wrong hypothesis about the inference capability of a passive adversary and results in false sense of security, which will expose the legitimate systems to severe threats with little awareness. In this project, we seek to understand the fundamental limits in passive inference of wireless channel characteristics, and further advance our knowledge and practice in wireless security.

The separation of control and data plane in SDN architectures helps merge packet and circuit paradigms into a single architecture and enables logical centralization of the control function. This enables new thinking about solutions to path optimization problems frequently encountered in networking, from routing to traffic engineering. The SERPENT project proposes to develop effective solutions for representing, storing and manipulating network state using rich semantic models such that path and topology embedding problems can be solved using a semantic database framework. This will simplify creation of novel network control and management systems able to cope with increasingly complex user requirements.

Achieving scalable processing of the increasing amount of publicly-available Semantic Web data will hinge on parallelization. The Map-Reduce programming paradigm recently emerged as a de-facto parallel data processing standard and has demonstrated effectiveness with respect to structured and unstructured data. However, Semantic Web data presents challenges not adequately addressed by existing techniques due to its flexible, fine-grained data model and the need to reason beyond explicitly represented data. This project will investigate optimization techniques that address these unique challenges based on rethinking Semantic Web data processing on Map-Reduce platforms from the ground, up - from query algebra to query execution.

Knowledge graphs are information networks with a specific topology that can be modeled as algebraic data types in a type system called Flutes, created by Datanova Scientific to rigorously analyze formal approaches to semantic integration. This project will demonstrate the capability of Flutes typing for summarizing knowledge graphs, collecting and collapsing high-dimensional data into low-dimensional data. Typing will help optimize queries based on column values similar to NOSQL databases. By using Flutes to implement a tactical knowledge base for DoD and other agencies, front-end applications can avoid including code for data validation, checking integrity constraints, or mission data extraction

DO3 Task Order 2.9 KRM

Human-system interactions is an integral part of any system. Because the vast majority of ordinary users have limited technical knowledge and can easily be confused and/or worn out by repeated security notifications/questions, the quality of users? decisions tends to be very low. On the other hand, any system targeting end-users must have the flexibility to accommodate a wide spectrum of different users, and therefore needs to get the full range of users involved in the decision making loop. This dilemma between fallible human nature and inevitable human decision making is one main challenge to the goal of improving security. In this project, we aim at developing principles and mechanisms for usable risk communication and control. The major technical innovations include (1) multi-granularity risk communications; (2) relative risk information in the context of comparison with alternatives; (3) Discover and integrate risk information from multiple sources; (4) Expand opportunities for risk communication and control.

This project develops a high-level notion of context that exploits the capabilities of next genera-tion networks to enable applications that deliver better user experiences. In particular, it exploits mobile devices always with a user to capture key elements of context: the user's location and, through localization, characteristics of the user's environment.

We propose to develop techniques that will enable humans to train computers efficiently and intuitively. In this proposed work, we draw inspiration from the ways that human trainers teach dogs complex behaviors to develop novel machine learning paradigms that will enable intelligent agents to learn from human trainers quickly, and in a way that humans can intuitively take advantage of. This research aims to return to the basics of programming---it seeks to develop novel methods that allow humans to tell computers what to do. More specifically, this research will develop learning techniques that explicitly model and leverage the implicit communication channel that humans use while training, a process akin to interpreting the pragmatic implicature of a natural language communication. We will develop algorithms that view the training process as an intentional communicative act, and can vastly outperform standard reward-seeking algorithms in terms of the speed and accuracy with which human trainers can generate desired behavior.

We propose to develop tools and techniques that will enable more effective two-way communication between dogs and handlers. We will work to create non-invasive physiological and inertial measuring devices that will transmit real-time information wirelessly to a computer. We will also develop technologies that will enable the computer to train desired behaviors using positive reinforcement without the direct input from humans. We will work to validate our approach using laboratory animals in the CVM as well as with a local assistance dog training organization working as a consultant.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

Optical networking forms the foundation of the global network infrastructure, hence the planning and design of optical networks is crucial to the operation and economics of the Internet and its ability to support critical and reliable communication services. With this research project we aim to make contributions that will lead to a quantum leap in the ability to solve optimally a range of optical design problems. In particular, we will develop compact formulations and solution approaches that can be applied efficiently to instances encountered in Internet-scale environments. Our goal is to lower the barrier to entry in fully exploring the solution space and in implementing and deploying innovative designs. The solutions we will develop are "future-proof" with respect to advances in DWDM transmission technology, as the size of the corresponding problem formulations is independent of the number of wavelengths.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

One of the most significant advances for large-scale scientific simulations has been the advent of Adaptive Mesh Refinement, or AMR. By using dynamic gridding, AMR can achieve substantial savings in memory, computation, and disk resources while maintaining or even increasing simulation accuracy, relative to static, uniform gridding. However, the resultant non-uniform structure of the simulation mesh produced by AMR methods cause inefficient access patterns during data analysis and visualization. Given the exponential increase in simulation output, the massive I/O operations are becoming a substantial bottleneck in simulations and analysis. To efficiently analyze AMR data, we propose an integrated, three-prong approach that aims: (a) To devise an AMR query model; (b) To develop effective in situ indexing and query processing methods for AMR data analytics; and (c) To investigate data storage layout strategies for AMR data retrieval optimized for analytics-induced heterogeneous data access patterns. The results, algorithms, and software will be in the public domain.

Dr. Nagiza Samatova's joint work with NC State University and Oak Ridge National Laboratory (ORNL) will provide the interface between the two organizations aiming to collaboratively address computational challenges in the Scientific Data Management, and the Large-Scale Analysis of DOE-mission applications. (Supplement)

Accelerating the rate of insight and scientific productivity demands new solutions to managing the avalanche of data expected in extreme-scale. Our approach is to use tools that can reduce, analyze, and index the data while it is still in memory (referred to as "in-situ" processing of data). ). In order to deal with the large amount of data generated by the simulations, our team has partnered with many application teams to deliver proven technology that can accelerate their knowledge discovery process. These technologies include ADIOS, FastBit, and Parallel R. In this proposal we wish to integrate these technologies together, and create a runtime system that will allow scientist to create an easy-to-use scientific workflow system, that will run in situ, in extra nodes on the system, which is used to not only accelerate their I/O speeds, but also to pre-analyze, index, visualize, and reduce the overall amount of information from these solutions.

LAS DO5 Samatova Task 5.2 Instrumentation

Dr. Nagiza Samatova's joint work with NC State University and Oak Ridge National Laboratory (ORNL) will provide the interface between the two organizations aiming to collaboratively address computational challenges in the Scientific Data Management, Data-Intensive Computing for Understanding Complex Biologicial Systems, Knowledge Integration for the Shewanella Federation, and the Large-Scale Analysis of Biologicial Networks with Applications to Bioenergy Production.

NCSU through the SAS GA will provide research and analysis to SAS as set forth in this Agreement. Such research and analysis shall include, but is not limited to, research, generation, testing, and documentation of operations research software. SAS GA will provide such services for SAS' offices in Cary, North Carolina, at such times as have been mutually agreed upon by the parties.

The SDAV is a unique and comprehensive combination of scientific data management, analysis, and visualization expertise and technologies aimed at enabling scientific knowledge discovery for applications running on state-of-the-art computational platforms located at DOE's primary computing facilities. This integrated institute focuses on tackling key challenges facing applications in our three focus areas through a well-coordinated team and management organization that can respond to changing technical and programmatic objectives. The proposed work portfolio is a blend of applied research and development, aimed at having key software services operate effectively on large distributed memory multi-core, and many-core platforms and especially DOE's open high performance computing facilities. Our goal is to create an integrated, open source, sustainable framework and software tools for the science community.

Physical science applications such as nanoscience, fusion science, climate and biology generate large-scale data sets from their simulations and high throughput technologies. This necessitates scalable technologies for processing and analyzing this data. We plan to research and develop advanced data mining algorithms for knowledge discovery from this complex, high-dimensional, and noisy data. We will apply these technologies to DOE-mission scientific applications related to fusion energy, bioenergy, understanding the impacts of climate extremes, and insider threat detection and mitigation.

The goal is to provide a computational capability for effective and efficient exploration of high-resolution climate networks derived from multivariate, uncertain, noisy and spatio-temporal climate data. We plan to increase the efficiency and climatologically relevancy of the network patterns identification through integrated research activities focused on: (a) supporting comparative analysis of multiple climate networks; (b) constraining the search space via exploiting the inherent structure (e.g., multi-partite) of climate networks; (c) establishing the foundation to efficiently update solutions for perturbed (changing) graphs; and (d) designing and implementing parallel algorithms scalable to thousands of processors on multi-node multi-core supercomputer architectures.

NC State University, in partnership with University of Michigan, Purdue University, University of Illinois at Urbana Champaign, Kansas State University, Georgia Institute of Technology, NC A&T State University, Los Alamos National Lab, Oak Ridge National Lab, and Pacific Northwest National lab, proposes to establish a Consortium for Nonproliferation Enabling Capabilities (CNEC). The vision of CNEC is to be a pre-eminent research and education hub dedicated to the development of enabling technologies and technical talent for meeting the grand challenges of nuclear nonproliferation in the next decade. CNEC research activities are divided into four thrust areas: 1) Signatures and Observables (S&O); 2) Simulation, Analysis, and Modeling (SAM); 3) Multi-source Data Fusion and Analytic Techniques (DFAT); and 4) Replacements for Potentially Dangerous Industrial and Medical Radiological Sources (RDRS). The goals are: 1) Identify and directly exploit signatures and observables (S&O) associated with special nuclear material (SNM) production, storage, and movement; 2) Develop simulation, analysis, and modeling (SAM) methods to identify and characterize SNM and facilities processing SNM; 3) Apply multi-source data fusion and analytic techniques to detect nuclear proliferation activities; and 4) Develop viable replacements for potentially dangerous existing industrial and medical radiological sources. In addition to research and development activities, CNEC will implement educational activities with the goal to develop a pool of future nuclear non-proliferation and other nuclear security professionals and researchers.

DO3 Task Order 2.6 Future States

Over the past ten years, lecture hall partitions have emerged as fundamental structures in combinatorics and number theory, leading to new generalizations and new interpretations of several classical theorems. This project takes a geometric view of lecture hall partitions and uses polyhedral geometry to investigate their remarkable properties.

Computer simulation is important for scientific research in many disciplines. Many such programs are complex, and transfer a large amount of data in a dynamically changing pattern. Memory performance is key to maximizing computing efficiency in the era of Chip Multiprocessors (CMP) due to the growing disparity between the slowly expanded memory bandwidth and the rapidly increased demands for data by processors. The importance is underlined by the trend towards exascale computing, in which, the processors are expected to each contain hundreds or thousands of (heterogeneous) cores. Unfortunately, today’s computer systems lack support for high degree of memory transfer. This project proposes to improve memory performance of dynamic applications by developing two new techniques that are tailored especially for the emerging features of CMP. The first technique is asynchronous streamlining, which analyzes the memory reference patterns of an application during runtime and regulates both control flows and memory references on the fly. The second technique is neighborhood-aware locality optimizations, which concentrates on the non-uniform relations among computing elements. This research will produce a robust tool for scientific users to enhance program locality on multi- and many-core systems that is not possible to achieve with existing tools. Further, it will contribute to the advancement of computational sciences and promote academic research and education in the challenging field of scientific computing.

In this project, Dr. Xipeng Shen and his team are building a new paradigm of program optimizations. It is motivated by a growing gap between trends in processor development and needs of modern data-intensive dynamic applications. This class of applications, ranging from differential equation solvers to data mining tools to particle dynamics simulations, play an essential role in science and humanity. But they feature tremendous data accesses and complex patterns in data accesses or control flows. The properties make them a great challenge for modern processors which are evolving exactly opposite to these applications' needs: A chip's aggregate computing power is rapidly outgrowing memory bandwidth; the rise of throughput-oriented manycores makes system throughput even more sensitive to irregular computations. The paradigm being built by Dr. Shen's team, namely "non-uniformity--centric optimizations", distinctively takes the non-uniform inter-core relations in modern systems as the first-order constraint for program optimizations. Specifically, they are developing a framework named PipeReg as a new way to reorganize data accesses and threads during run time to reduce the influence of irregular computations on the throughput of massively parallel processors. Meanwhile, they are investigating a novel kind of program transformations called neighborhood-aware transformations, which exploits the non-uniform interactions among threads in on-chip storage (e.g., shared cache) in a multi-socket multicore system. Together, the two techniques will synergistically remove some important barriers for data-intensive dynamic applications to tap into the full power of future computing systems. The outcome from this research will provide essential support for enhancing the computing efficiency of data-intensive dynamic applications in the era of heterogeneous parallel systems. Because of the critical roles of these applications, this research will help foster sustained advancement in science, commerce, health, and so on. Beyond its technical content, this project stresses technology transfer, develops new teaching materials and tools, emphasizes demographic diversity, and improves the training of both graduate and undergraduate students.

By analyzing and predicting program dynamic behaviors, program behavior analysis offers the fundamental support for program transformations and resource management. Its effectiveness is crucial for the maximization of computing efficiency. This research proposes to include program inputs---a so far virtually ignored dimension---into the focus of program behavior analysis, cultivating a new paradigm, namely input-centric program behavior analysis and adaptation. This input-centric paradigm will create many new opportunities for enhancing the matching between software and hardware, hence significantly improving the performance and power efficiency in modern computing. The proposed technique, input-centric program behavior analysis and adaptation, consists of three components. The first two components, program input characterization and input-behavior modeling, resolve the complexities of program inputs, extract important features, and recognize the correlations between characterized input features and program behaviors. The third component, input-centric adaptation, capitalizes on the novel opportunities that the first two components create, making dynamic optimizations proactive and holistic, but without losing the adaptivity to inputs and environmental changes. Together, the three components make evolvable programming systems more feasible than before. In such a system, the input-behavior models embody the central knowledge base, which grows incrementally across program production runs. As the knowledge base becomes larger, behavior prediction becomes more accurate, stimulating better software-hardware matching and making the program and runtime systems perform increasingly better. Because of the fundamental role of program behavior analysis in software-hardware matching, this research helps pave the way for advancing the optimizations in various layers in the software execution stack (compilers, virtual machines, OS, etc.).

A component (e.g., a function or loop) in a program often exhibits different behaviors (e.g., execution paths) in a different context. Such a context sensitivity exists in High Performance Computing (HPC) applications, and even more commonly in Business Analytics and Optimizations (BOA) programs. This collaboration with IBM aims at developing context-aware correlation-based program optimizations, a new way to tackle context sensitivity in code specializations that effectively removes some limitations in current compiler technology.

In this project, we propose to build up context-aware correlation-based program optimizations, a new way to tackle context sensitivity in code specializations that effectively removes some limitations in current compiler technology.

Contemporary architectures are adopting an integrated design of conventional CPUs with accelerators on the same die with access to the same memory, albeit with different coherence models. Examples include AMD's Fusion architecture, Intel's integrated main-stream CPU/GPU product line, and NVIDIA Tegra's integrated graphics processor family. Integrated GPUs feature shared caches and a common memory interconnect with multicore CPUs, which intensify resource contention in the memory hierarchy. This creates new challenges for data locality, task partitioning and scheduling, as well as program transformations. Most significantly, a program running on GPU warps and CPU cores may adversely affect performance and power of one another. The objective of this work is to understand these novel implications of fused architectures by studying their effects, qualifying their causes and quantifying the impacts on performance and energy efficiency. We propose to advance the state-of-the-art by creating spheres of isolation between CPU and GPU execution via novel systems mechanisms and compiler transformations that reduce cross-boundary contention with respect to shared hardware resources. This synergy between systems and compiler techniques has the potential to significantly improve performance and power guarantees for co-scheduling pgrams fragments on fused architectures. impact: The proposed work, if successful, has the potential to transform resource allocation and scheduling at the systems level and compiler optimizations at the program level to create a synergistic development environment with significant performance and power improvements and vastly increased isolation suitable for synergistic co-deployment of programs crossing boundaries on innovative fused architectures.

Intelligent machines are purported to be the back-bone of the cybermanufacturing initiative.Yet, the conventional approach to making a machine ‘cyber-enabled’, is to outfit the machine with an array of multi-modal sensors which is then integrated to the network and enterprise system through communication and computing platforms. To make further development challenging, almost all industrial machine vendors have closed hardware and software architecture which makes it difficult for extensibility and adaptation to a cyber-manufacturing environment. We propose a new architecture, which we term as the – ‘Industrial Machine Operating System - iMOS’, will be a flexible framework for writing machine software. It will be a collection of hardware configurations, data structures, tools, libraries and semantics to simplify the task of creating a cyber-physical enabled manufacturing machine, designed to operate across a wide variety of manufacturing process platforms. (Total Award Amount $299,999.00

LAS DO5 Singh Task 5.2 Instrumentation

DO3 Task Order 2.11 Mission Enabling

DO3 Task Order 2.8 Analytic Workflow

DO3 Task Order 2.9 KRM

This project will develop policy-based service governance modules for the Oceanographic Observatories Initiative (OOI) Cyberinfrastructure. The main objectives of the proposed project include (1) formulating the key conceptual model underlying the patterns of governance; (2) formalizing "best practices" patterns familiar to the scientific community and seeding the cyberinfrastructure with them; (3) understanding user requirements for tools that support creating and editing patterns of governance

LAS DO5 St. Amant Task 5.4 Informatics

DO3 Task Order 2.5 Cognitive Processing

The Institute of Next Generation IT Systems (ITng) recently completed Phase I of Law Enforcement Data Fusion for Northrup Grumman (NG). Activities resulted in the collection of multiple sources of data concerning possible surrogate markers for predicting crime. Unstructured text analytics were performed that demonstrated the feasibility of this technology in isolating facts around crimes that appear in news and social media. In the second phase of the project, ITng along with select faculty from the College of Computer Science (as required) will enhance the model by incorporating streaming data, machine learning, and a web interface for data access.

This research proposal is aiming to focus on embracing data fusion through the aggregation and disambiguation of multiple data streams that allow for filtering data with it’s subsequent representation. Data types include structured data, structured metadata, and unstructured text. There is no harmonization of data types. The process of disambiguation will be performed using an unstructured text analytics to extract evidence from documents using rules-based extraction coupled with statistical methods to improve rule generation. Structured data sets will be harmonized through aggregation and recombination using H-based systems. Subsequent data sets will be deployed into schemas that allow for rapid assessment data sets for report building (queries). The research will expand to investigate data fusion integrated with machine learning techniques that can add a higher level of data selectivity.

LAS DO5 Task Streck 5.2 Instrumentation

DO3 Task Order 2.2 Infrastructure

The PI's unique combination of expertise in structural graph theory and scalable graph algorithms for data-driven science is necessary to ensure the success of ORNL-based projects using applied discrete mathematics to enable advances in graph analysis, anomaly detection, cybersecurity, quantum computing, and computational science. The PI will direct and conduct fundamental research, collaborate with ORNL staff, write up research results for peer-reviewed publication, give presentations, and mentor students and junior staff as appropriate.

This extension to the PARSiNG project focuses on issues related to improving accessibility and usability for downstream analysts of the related open source software toolkit. This may include new features (such as a graphical interface), improved I/O and data format support, extension of the modular framework to additional DARPA-relevant problems, and testing and incorporation of algorithmic coloring advances.

"Big Data" sources for many real-world applications pose numerous challenges to understanding the complex and possibly hidden relationships between components of a complex network. Furthermore, these networks often consist of heterogeneous entities and types of relationships, and many existing algorithms for computing network features and similarity are not directly applicable. In order to draw actionable insights, analysis need to identify events of interest, place them in context, and understand their impact. Existing approaches which emphasize visualization (at the expense of analytics), struggle to coherently present networks with hundreds of entities, whereas practical applications require hundreds of thousands (or more). We propose to develop methods which integrate ideas from graph theory with multi-scale modeling (since events of interest may occur at different levels of granularity/contexts within the data) to improve comprehension of such relational data and form a foundation for novel methods of visualization and interaction.

The PI's unique combination of expertise in structural graph theory and scalable graph algorithms/big data is necessary to ensure the success of ORNL-based projects using applied discrete mathematics to enable advances in graph generation and HPC benchmarking, social network analysis, and multi-scale graph decompositions for the Department of Energy and Department of Defense. The PI will direct and conduct fundamental research, collaborate with ORNL staff, write up research results for peer-reviewed publication, give presentations, and mentor students and postdoctoral staff as appropriate.

Understanding and identifying intermediate-scale structure is key to designing robust tools for data analysis, just as the interdependence of local interactions and global behavior is key in many science domains. We thus focus on constructing a theory and tools for using this structure to improve analysis and identification of relationships in massive graph data. Through careful integration of tools from graph theory, computational complexity, statistics, and parallel algorithm design, the proposed work will derive novel measures of graph similarity based on structural representations and application-inspired features of interest. We will design efficient, scalable sampling algorithms which leverage inherent sparsity and structure to de-noise and improve accuracy of parameter estimation. As a specific example of science domain impact, we focus on improving understanding of the brain. Applying our new tools for characterizing graph-theoretic structure in such networks, scientists will be able to build higher fidelity models of brain network formation and evolution. Additionally, efficient algorithms from the associated parameterized framework will enable rapid comparison of regions and identification of discrepancies, abnormalities, and influential components for specific tasks.

NCSU graduate student will develop a machine learning approach to linking extreme atmospheric ridging events with extreme surface temperatures, employing a Gaussian Process (GP)-based predictive analysis tool that leverages the predictive value of spatial and temporal correlations and builds on ORNL’s past success in spatial classification, temporal change prediction, and parallelizing GP for large spatiotemporal extents.

Pattern discovery and predictive modeling from seemingly related “Big Data” represented as massive, ad-hoc, heterogeneous networks (e.g., extremely large graphs with complex, possibly unknown structure) is an outstanding problem in many application domains. To address this problem, we will be designing graph-mining algorithms capable of: (i) discovering relationship-patterns from such data and (ii) use those discovered patterns as features for classification and predictive modeling. In particular, we will develop, implement, and demonstrate novel, automated and scalable graph-pattern discovery algorithms on two different architectures: (1) Cray’s Urika, and (2) Hadoop based Cluster. The research will bring forth the advantages and disadvantages these two different architectures and shows possible optimization approaches for pattern discovery from big data.

NCSU graduate student will develop a machine learning approach to linking extreme atmospheric ridging events with extreme surface temperatures, employing a Gaussian Process (GP)-based predictive analysis tool that leverages the predictive value of spatial and temporal correlations and builds on ORNL’s past success in spatial classification, temporal change prediction, and parallelizing GP for large spatiotemporal extents.

Pattern discovery and predictive modeling from seemingly related “Big Data” represented as massive, ad-hoc, heterogeneous networks (e.g., extremely large graphs with complex, possibly unknown structure) is an outstanding problem in many application domains. To address this problem, we will be designing graph-mining algorithms capable of: (i) discovering relationship-patterns from such data and (ii) use those discovered patterns as features for classification and predictive modeling. In particular, we will develop, implement, and demonstrate novel, automated and scalable graph-pattern discovery algorithms on two different architectures: (1) Cray’s Urika, and (2) Hadoop based Cluster. The research will bring forth the advantages and disadvantages these two different architectures and shows possible optimization approaches for pattern discovery from big data.

Develop change detection techniques for: (1) biomass monitoring via MODIS time series analysis, (2) Provide expert services to parallelize Gaussian Process (GP) learning on GPUs, and (3) implement a multi-resolution object-based change analysis framework for detecting urban sprawl using very high resolution satellite imagery.

Since 1995, the Triangle Computer Science Distinguished Lecturer Series (TCSDLS) has been hosting influential university researchers and industry leaders from computer-related fields as speakers at the three universities within the Research Triangle Area. The lecturer series, sponsored by the Army Research Office (ARO), is organized and administered by the Computer Science departments at Duke University, NC State University, and the University of North Carolina at Chapel Hill. This proposal argues for continuation, for an additional 3 years, of this highly successful lecturer series which is being led by Duke University.

In partnership with industry and faculty from across the country, this project will develop a transformative approach to developing the communication abilities (writing, speaking, teaming, and reading) of Computer Science and Software Engineering students. We will integrate communication instruction and activities throughout the curriculum in ways that enhance rather than replace their learning technical content and that supports development of computational thinking abilities of the students. We will implement the approach at two institutions. By creating concepts and resources that can be adapted by all CS and SE programs, this project also has the potential to increase higher education's ability nationwide to meet industry need for CS and SE graduates with much better communication abilities than, on average, is the case today. In addition, by using the concepts and resources developed in this project, CS and SE programs will be able to increase their graduates' mastery of technical content and computational thinking.

Jamming resistance is crucial for applications where reliable wireless communication is required, such as rescue missions and military applications. Spread spectrum techniques such as Frequency Hopping (FH) and Direct Sequence Spread Spectrum (DSSS) have been used as countermeasures against jamming attacks. However, these anti-jamming techniques require that senders and receivers share a secret key to communicate with each other, and thus are vulnerable to insider attacks where the adversary has access to the secret key. The objective of this project is to develop a suite of techniques to defend against insider jammers in DSSS and FH based wireless communication systems. We will develop novel and efficient insider-jamming-resistant techniques for both DSSS- and FH-based wireless communication systems. Our proposed research consists of two thrusts. The first thrust is to develop novel spreading/despreading techniques, called DSD-DSSS (which stands for DSSS based on Delayed Seed Disclosure), to enhance DSSS-based wireless communication to defend against insider jamming threats, while the second thrust is to develop a new approach, called USD-FH (which stands for FH based on Uncoordinated Seed Disclosure), to enable sender and receivers using FH to communicate without pre-establishing any common secret hopping pattern. A key property of our new approaches is that they do not depend on any secret shared by the sender and receivers. Our solution has the potential to significantly enhance the anti-jamming capability of today?s wireless communication systems.

The proposed work examines novel services built on top of public cloud infrastructure to enable cost-effective high-performance computing. We will explore the on-demand, elastic, and configurable features of cloud computing to complement the traditional supercomputer/cluster platforms. If successful, this research will result in tools that adaptively aggregate, configure, and re-configure cloud resources for different HPC needs, with the purpose of offering low-cost R&D environments for scalable parallel applications.

Simulating realistic mobility patterns of mobile devices is important for the performance study of mobile networks because deploying a real testbed of mobile networks is extremely difficult, and furthermore, even with such a testbed, constructing repeatable performance experiments using mobile devices is not trivial. Humans are a big factor in simulating mobile networks as most mobile nodes or devices (cell phones, PDAs and cars) are attached to or driven by humans. Emulating the realistic mobility patterns of humans can enhance the realism of simulation-based performance evaluation of human-driven mobile networks. Our NSF-funded research that ends this year has studied the patterns of human mobility using GPS traces of over 100 volunteers from five different sites including university campuses, New York City, Disney World, and State Fair. This research has revealed many important fundamental statistical properties of human mobility, namely heavy-tail flight distributions, self-similar dispersion of visit points, and least-action principle for trip planning. Most of all, it finds that people tend to optimize their trips in a way to minimize their discomfort or cost of trips (e.g., distance). No existing mobility models explicitly represent all of these properties. Our results are very encouraging and the proposed research will extend the work well beyond what has been accomplished so far. . We will perform a measurement study tracking the mobility of 100 or 200 students in a campus simultaneously, and analyze the mobility patterns associated with geo-physical and social contexts of participants including social networks, interactions, spatio-temporal correlations, and meetings. . We will cast the problem of mobility modeling as an optimization problem borrowing techniques from AI and Robotics which will make it easy to incorporate the statistical properties of mobility patterns commonly arising from group mobility traces. The realism of our models in expressing human mobility will surpass any existing human mobility models. . We will develop new routing protocols leveraging the researched statistical properties found in real traces to optimize delivery performance. The end products of the proposed research is (a) a new human mobility model that is capable of realistically expressing mobility patterns arising from reaction to social and geo-physical contexts, (b) their implementation in network simulators such as NS-2/3 and GloMoSim, (c) mobility traces that contain both trajectories of people in a university campus and contact times, (d) new efficient routing protocols for mobile networks

RECOGNIZE is a visual quiz game, with a mechanic much like the TV game show “Name That Tune.” Rather than naming a snippet of music, players match a slowly revealed “source” image (e.g., a picture of the artist Salvador Dali) to one of several “target” images (a different picture of Dali, as well as several other artists). RECOGNIZE is unique in its fully visual quizzing mechanic, with both questions and answers posed visually. This project will further develop RECOGNIZE into a viable commercial product with a broad range of educational applications, including individual instruction, group exercises, and distance learning.

According to a 2010 report that was based on the interviews from 2,800 Information Technology professionals worldwide, the gap between hacker threats and suitable security defenses is widening, and the types and numbers of threats are changing faster than ever before . In 2010, Jim Gosler, a fellow at the Sandia National Laboratory who works on countering attacks on U.S. networks, claimed that there are approximately 1,000 people in the country with the skills needed for cyber defense. Gosler went on to say that 20 to 30 times that many are needed. Additionally, the Chief Executive Officer (CEO) of the Mykonos Software security firm indicated that today's graduates in software engineering are unprepared to enter the workforce because they lack a solid understanding of how to make their applications secure. Particularly due to this shortage of security expertise, education of students and professionals already in the workforce is paramount. In this grant we provide a plan for motivating and providing software security education to students and professionals.

Touch interfaces on mobile phones and tablets are notoriously error prone in use. One plausible reason for slow progress in improving usability is that research and design efforts in HCI take a relatively narrow focus on isolating and eliminating human error. We take a different perspective: failure represents breakdowns in adaptations directed at coping with complexity. The key to improved usability is understanding the factors that contribute to both expertise and its breakdown. We propose to develop cognitive models of strategies for touch interaction. Our research will examine the detailed interactions between users perceptual, cognitive, and motor processes in recognizing, recovering from, and avoiding errors in touch interfaces. Our proposal is for three stages of research: exploratory experiments, analysis and modeling, and finally validation experiments.

Critical cyber systems must inspire trust and confidence, protect the privacy and integrity of data resources, and perform reliably. Therefore, a more scientific basis for the design and analysis of trusted systems is needed. In this proposal, we aim to progress the Science of Security. The Science of Security entails the development of a body of knowledge containing laws, axioms and provable theories relating to some aspect of system security. Security science should give us an understanding of the limits of what is possible in some security domain, by providing objective and quantifiable descriptions of security properties and behaviors. The notions embodied in security science should have broad applicability - transcending specific systems, attacks, and defensive mechanisms. A major goal is the creation of a unified body of knowledge that can serve as the basis of a trust engineering discipline, curriculum, and rigorous design methodologies. As such, we provide eight hard problems in the science of security. We also present representative projects which we feel will make progress in the discipline of the science of security.

Since August 2011, North Carolina State University (NCSU) analytics-focused Science of Security Lablet (SOSL) has embraced and helped build a foundation for the NSA vision of the Science of Security (SoS) and a SoS community. Jointly with other SOSLs, we formulated five SoS hard problems, which lie at the core of the BAA. At NCSU, data-driven discovery and analytics have been used to formulate, validate, evolve, and solidify security models and theories as well as the practice of cyber-security. We propose to (1) investigate solutions to five cross-dependent hard problems, building on our extensive experience and research, including in the current SOSL; (2) advance our SoS community development activities; and (3) enhance our evaluation efforts regarding progress on the hard problems by bringing in experts on science evaluation.

Virtual cinematography, the use of a virtual camera within a three dimensional virtual world, is being used increasingly to communicate both in entertainment contexts as well as serious ones (e.g., training, education, news reporting). While many aspects of the use of virtual cameras are currently automated, the control of the camera is currently determined using either a pre-programmed script or a human operator controlling the camera at the time of filming. This project seeks to develop a model of virtual cinematography that is both computational -- providing a software system capable of generating camera control directives automatically -- and cognitive -- capable of modeling a viewer's understanding of an unfolding cinematic.

LAS DO5 Young Task 5.3 Sensemaking

This proposal would fund the research and design phase of building the ALICE engine. The project will focus on creating a conceptual model for how to build an adaptive learning space; an architecture for the Artificial Intelligence (AI) engine and technology core; a series of "proof of concept" functional prototypes for collecting data and creating content; and a continuous assessment model for measuring the success of the AI as far as the quality of its engagement with the community and the success of the engine at enhancing a given technology-rich research and learning space.

DO3 Task Order 2.4 Narrative Processing