Javed Mostafa - US grants

Abstract

IIS-9817572
Palakal, Mathew
Indiana University
$101,604 - 12 mos.

DLI Phase 2: A Distributed Information Filtering System for Digital Libraries

This is the first year funding of a three year continuing award. The proposed research is aimed at designing and developing a distributed intelligent information distribution and filtering system that provides personalized information services to the user while minimizing direct user involvement. The system is intended to traverse the internet to retrieve the most relevant information of interest to the user. Information filtering will be realized using a information agents, and will involve integration of advanced concepts and techniques from the domains of artificial intelligence, information retrieval, and distributed object computing. The agents will contain models of network-based dynamic information resources and will have the capability to learn changing patterns of an individual user's interest.

Four key basic research areas to be addressed are: methods for adapting various knowledge structures associated with an agent
new and robust agent architectures
agent collaboration protocols based on a natural or artificial economic framework
agent-driven information service operations

The explosive growth of biological information sources, available over the Internet, has given rise to both opportunities and challenges for biological and medical researchers. The opportunities they provide are both scientific (e.g., understanding the information encoded in elementary biological structures) as well as technological (e.g., new drug discovery). The challenges, on the other hand, lie in how to efficiently discover, among the vast volume of information, the items that are relevant or interesting to a given researcher. The objective of the proposed research is to investigate related basic research problems and develop a biological information delivery system in a collaborative project between computer scientists, information scientists, and biological researchers. The specific plans include developing methods to make the proposed system pro-active (surveying evolving on-line sources for relevant information), personalized (cognizant of a particular researcher's interests), adaptive (able to react to changes in the information sources as well as user interests or objectives), and capable of integrating multi-format data. The impact of this research is a significant enhancement in the ability of students and researchers in biological sciences to efficiently utilize on-line resources, while generating methods for computerized analysis of biological data and providing computerized support for new scientific discovery.

This project is applying advances in digital library (DL) technologies to the emerging domain of bioinformatics, and developing interaction tools that support learning based on identifying and visualizing associations among key dimensions of bioinformatics resources. The project is making these current resources, which are mainly utilized by expert biologists, available to bioinformatics students. The project addresses issues regarding: the use of a wide variety of formats and representations to store bioinformatics information; the application of DL technologies particularly in the realm of data description and exchange; mapping of metadata associated with bioinformatics information to data description standards compatible with DL technologies; and building clients that take advantage of data dissemination protocols such as the Open Archives Initiative in order to support novel browse, search, and analysis functions based on visualizations. The project team is also integrating DL and Grid computing technologies, and utilizes bioinformatics resources of Indiana University such as the Drosophila Genome Flybase, the IUBio Archive that contains euGenes eukaryote genes data, and the Bionet news archive and related software.

This is a collaborative grant with two PIs; Javed Mostafa of Indiana, and David Stark of Columbia.

Intellectual Merit
With a grant from the NSF Digital Government Program, David Stark has been studying the role of information technologies in the public debate surrounding the rebuilding of Lower Manhattan in the wake of the September 11 attacks on the World Trade Center. In the process of conducting that research Stark's team has assembled an extensive digital archive containing 5,000 participant oral statements from one town hall meeting and an additional 19,000 oral statements collected at 240 different venues around New York City in the 'Imagine New York Envisioning Workshops'. These gathered statements provide a rich opportunity for testing various strategies of computer-assisted interpretation because they provide an opportunity to compare the conceptual patterns discerned by human intelligence with findings reached through the analytical methods of artificial intelligence. Supporting the initial explotation of that archive is the purpose of this grant.

The technical component of this grant arises from work Javed Mostafa has done under an NSF ITR grant. Data mining research concentrating on spontaneous human conversations is at an early stage of development. Mostafa's approach to data mining can offer different ways to analyze the same data. The project has three specific goals: 1) to detect emergent concepts by applying techniques that do not impose any a priori conditions; 2) to use techniques for analyzing known concepts by applying constraints on the mining process, and 3) to develop visualization of the results to facilitate interpretation by social scientists and support direct validation by citizen participants.

Broad Impact
Computer mediated communication offers new channels for citizens to express their views to elected officials and government agencies. Often, the resulting deluge of comments poses a technical and political challenge. How can officials/agencies make sense of large-scale citizen input? How can meaningful patterns be efficiently and effectively identified? This project will contribute to advancing understanding of the opportunities and the limitations of computer-assisted interpretation. Its findings will be of considerable interest to scholars as well as to government managers responsible for the rebuilding of lower Manhattan.

Summary
Many challenges involved in creating new data mining tools demands an interdisciplinary collaboration for access to new data; this project offers such an opportunity. This time-critical testing of artificial intelligence methods will be important in understanding the public input to rebuilding lower Manhattan.

Consumers of health information have access to increasingly large numbers of sources from which to gather this information. Privacy preserving access to high quality health information will be an important adjunct to personal healthcare management. This project is developing a "patient-centered" approach to handle search complexity and provide personalized access to high quality health information while preserving the patient's privacy at all times. This research brings together an interdisciplinary team of computer and information scientists and health experts to develop a test-bed and simulation environment to support research in this emerging area.

The broader impact of the research will be twofold. Locally Indiana University offers a graduate degree in Health Informatics. This project will help inform and educate the next generation of health informaticians. The larger health informatics community will benefit from the availability of a test-bed to advance information retrieval in this important area.

Although digitization of health information and IT applications in health care have received significant focus in recent years (mainly from government officials interested in improving care while achieving more cost efficiencies), the concerns of patients and their needs, as they relate to information access, have not received a similar level of attention. There is a gap in expectation with regard to the role of health information systems between the general public, who are increasingly becoming more computer- and Web savvy, and health IT professionals. A deeper consideration of the gap is expected to help clarify the challenges and point to potential means for developing effective information systems that deliver health information.

The main goal of this national workshop is to focus on key challenges involved in delivering high quality digital health information directly to citizens and health care providers. The challenges cover many areas encompassing access, interpretation, decision-making, security/privacy, evaluation, and economic issues. A goal of the workshop is to create a forum where multi-disciplinary perspectives are supported to develop a scope for research in this area, identify critical research topics, and establish associated knowledge resources that may be helpful.

A critical impediment to taking full advantage of electronic health record (EHR) systems is the human computer interface (HCI). The project's core aim is to design, implement, and evaluate several new HCI features for clinical decision support (CDS) systems that have the potential for directly impacting point-of-care treatment and improving care for patients suffering from long-term illness. Some of the primary HCI features proposed include: 1) Visually enhanced interactive functions for EHR systems such as treatment trend maps and data indicators, 2) Patient data summary of critical data elements in a dashboard format, 3) Access to evidence-based recommendations relevant to individual patients, and, 4) Display of co-morbid conditions explicitly to encourage review and analysis in the context of routine patient care. The proposed work will explore the use of visual analytics to process patient and comparative population medical profile data to generate data summaries in the form of interactive visualizations (data views). The data views will be designed to provide physicians with the capability to filter patient data based on key data elements such as medications and co-morbid conditions, and to compare patient treatment trends with established evidence-based guidelines to identify treatment gaps and options. As part of the usability evaluation, the proposed work will engage end-users (i.e., care providers) directly in requirement generation, implementation, evaluation, and refinement of the CDS HCI. The broad impact of the proposed work include disseminating the outcome and findings through the network of Clinical Translational Science Awardees (60 academic medical centers) that are involved in developing novel informatics solutions to accelerate the translation of medical treatments and prevention strategies. Additionally, the MindsEye software will be distributed with appropriate documentation to contribute to health IT education and training and to spur practical IT engagements and new research activities.

The project aims to improve how humans, particularly those with degenerative cognitive conditions or limited cognitive capacities, interact with complex computing or computer-enabled services (broadly called cyber-systems). To advance the state of user interfaces (UI) and create more adaptive interaction modalities, the foundational knowledge regarding human-cyber interaction (HCI) must incorporate more neuro-physiological evidence. The project proposes a three-step plan for attaining this advance. First, three experimental subject groups will be carefully selected and recruited for comparative analysis: 1) healthy and young adults 2) demented or near-demented older adults, and 3) adults suffering from mild brain trauma. Second, experimental sessions will be conducted based on incrementally complex information searching and messaging tasks that require executing the tasks inside an MRI machine.

The goal of the experimental phase is to collect neuronal activation patterns using fMRI, as well as behavioral data associated with response time and accuracy. Third, the experimental findings will be analyzed to establish potential association among neuronal evidence, behavior, and UI performance. The broad technical aim of the initiative is to establish and refine methods for gathering neuro-physiological evidence under complex HCI conditions, and develop new user modeling techniques for supporting flexible and effective interaction. Development of advanced user interfaces capable of monitoring and establishing risk factors associated with impending or existing brain-degenerative conditions is a longer term translational aim. The resources and outcomes produced will be broadly shared among scholars in computer and information science disciplines, with the aim of promoting and supporting training of next-generation HCI researchers.

The proposed training program adds predoctoral and postdoctoral training to the Carolina Health Informatics Program (CHIP), which concentrates on both advancing research and training in biomedical and health informatics. CHIP is a truly interdisciplinary program co- created and co-sponsored by 1) School of Medicine, 2) Gillings School of Global Public Health, 3) School of Nursing, 4) Eshelman School of Pharmacy, 5) School of Dentistry, 6) School of Information & Library Science, and 7) Computer Science Department of the College of Arts & Sciences. The program is further strengthened by close partnerships with area health care organizations and companies specializing in biopharmaceuticals, heath care, IT and analytics. Recognizing the many challenges associated with health data analytics and use, the three foundational aspects of the CHIP training programs are: 1) Methods for health data analytics, 2) Building powerful interactive analytics systems that allow customization and support decision making for a wide variety of health care operations, and 3) Addressing translational challenges in implementing and managing information systems that leverage health data analytics effectively. Given that CHIP is a highly interdisciplinary program, it will draw upon its diverse faculty, research centers, laboratories, and other rich resources to develop three major types of activities as part of the proposed advanced predoctoral and postdoctoral training program. The three activities are: 1) Research methods and analytics techniques for clinical data science, imaging informatics, precision medicine, and big data analytics for population health, 2) Didactic learning based on courses in machine learning, biostatistics, big data management, and visual analytics and additional courses that are rooted in realistic health care problems, and 3) Large- scale, population-oriented system development, deployment, and outcome evaluation.

The scale and outcomes of undergraduate research in the US are poorly understood, in part because they are difficult to track and no standardized metrics exist for measuring them. While existing research has utilized surveys of graduates over time to obtain data on this topic, this method is both expensive and necessarily incomplete. Educational institutions and policymakers therefore rely on incomplete information when making policies that impact or promote undergraduate research activities. Given evidence that undergraduate research experiences impact students? willingness to undertake graduate study and careers in STEM, as well as continuing challenges to diversity and equity in these fields, more complete data is needed to inform both educational research and effective policymaking. This project addresses the problem by developing standards for measuring and understanding undergraduate research pathways using untapped data sources, including scientific publications by undergraduates and records maintained by the schools at which they study. The initial research dataset of undergraduate research products and outcomes focuses on undergraduate activities in multiple settings: research-focused universities, historically black colleges and universities (HBCUs), and STEM-focused companies. In addition to broadening the information base on undergraduate research available to educational institutions, policymakers, and researchers, this project advances new methods in educational data collection, data management, and dissemination. Preserving the privacy and security of educational records is extremely important and represents a technical challenge for researchers trying to collect and disseminate such data at scale. This project draws inspiration from previous innovations in health data systems that both preserve privacy and allow for responsible research on large-scale real-world data sets.

Project aims will be accomplished through the collaborative development and refinement of a data collection protocol, implementation of the data collection protocol across partner institutions, and initial development of a secure, compliant, and usable system for storing and working with data. A systematic literature review on methods for evaluating undergraduate research activities, outputs, and subsequent student outcomes will be conducted at the start of the project. Next, interviews will be conducted with appropriate staff at partner institutions to understand and evaluate their existing practices for tracking undergraduate research. Based on the information gathered through these steps, metrics to answer important questions about undergraduate research outcomes along with a corresponding data collection protocol will be jointly developed with partner institutions. An initial dataset will be collected using the developed protocol, focusing on undergraduate research in the fields of computer science and engineering. The policies, procedures, and research methods found to be most suitable to this work will be published to enable more effective and responsible research of undergraduate activity by the broader research community. Finally, requirements for a system to store and disseminate the data will be developed with the goal of providing a platform for future data collection and research at a national scale, across all undergraduate majors.

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