Past Grants/Awards

The COVID-19 pandemic has sparked thousands of new large-scale data projects globally. These COVID data infrastructures are essential: they enable the public, policymakers, public health officials, and others to see and comprehend particular aspects of the global health crisis. This research compares COVID data infrastructures in the U.S. and India, countries that share extremely high COVID infection rates as well as electoral democracy that encourages transparency; 'Data for Social Good' rhetoric; and large IT workforces. The project seeks to reveal how project leaders and contributors confront and manage the disruptions, hardships, and conflicts created by the pandemic. Working across different geographies and institutional settings, the research project will highlight how the pandemic impacts different communities in different ways. The research project will provide policymakers, technologists, and other leaders with insights and recommendations on how to improve the creation and maintenance of emergency data infrastructures. By understanding the dynamics of current COVID data infrastructures, we can be better prepared for the next emergency. This RAPID research project investigates the creation, maintenance, and real-time transformation of novel critical data infrastructures. It uncovers the debates, conflicts, orderings, and important decisions that shape and define COVID data-tracker systems. At a time when the pandemic is disrupting ongoing research across the globe, these data-trackers can provide insights into how to create and maintain resilient and sustainable research-enabling infrastructure under conditions of significant stress. This RAPID project uses cross-national comparative analysis of public COVID data projects in the U.S. and India in order to identify the key factors that enable data infrastructures to endure the social and material disruptions associated with the pandemic. The project's cross-national and comparative research design ensures that research findings are generalizable. COVID data infrastructures are dynamic: the information, practices, tools, and collaborators that populate these systems constantly evolve. Often, the important adaptations that shape critical data infrastructures are not easily preserved using current web archiving and cumulative public data preservation methods. Additionally, the project's research design will capture this otherwise ephemeral data--allowing the project to analyze and interpret how these infrastructures are created and maintained under adverse conditions. The project is informed by and will contribute to the scholarly literature on ethnographies of technology development, infrastructure studies, and crisis informatics. Research findings will support concrete recommendations for how these and future data infrastructure can be made (1) sustainable; (2) accountable to different publics; and (3) improved in order to help save lives.

The broader impact/commercial potential of this I-Corps project is the development of the explainable Artificial Intelligence (XAI) methods for healthcare data. Currently, the number of electronic medical records is increasing while machine learning and deep learning models, especially large language models, have been employed to address healthcare needs. However, the healthcare domain is highly regulated and explainability for the black-box AI model becomes increasingly critical for any AI application. Users need to comprehend and trust the results and output created by machine learning algorithms. The proposed XAI technology may be used to describe an AI model, its expected impact, and potential biases. Further, the proposed technology may be used to transfer AI predictions into explainable medical interventions to enable the last mile delivery of AI in healthcare The commercial potential of these technologies may impact three major groups: health insurance companies who may provide better care management interventions and achieve personalized care delivery based on XAI; health analytic companies who rely on explanation to further enhance their products and meet the government regulations; and medical device startups who demand explainable analytical outputs based on the collected data from medical devices to enrich their user experience.   This I-Corps project is based on the development of explainable Artificial Intelligence (XAI) methods applied to the healthcare industry. Providing explainability is critical for AI health applications. Healthcare is a unique domain with multimodality data: tableau data about patient demographic information, textual data about medical notes, time series data about vital sign measures, images about medical scan, and wavelet data about EEG and ECG. To provide a holistic view of these data, deep learning is used to create universal embeddings on different modalities of data and build the prediction models for health risks. But deep learning methods lack transparency and demand explainability. The proposed technology combines integrated gradients with ablation studies to identify the contributing factors of different data components in the explanation. In addition, the proposed platform adds knowledge graphs into the prediction and explanation workflow to detect the relationships between contributing features to generate an explanation with a holistic view, and translates weights or feature importance into risk scores to enable the last mile delivery of AI in healthcare. The proposed XAI method may be used to explain the importance of input data components, identify the contributing features at the individual patient level and the patient cohort level; scale and save computational resources; and self-improve by using reinforcement learning to enhance positive feedback.

The team has completed focus groups with iSchool Inclusion Institute participants where we piloted interview questions, tested and adjusted the questions, and gathered preliminary information on how community and belonging are cultivated. During the pandemic, we pivoted to longitudinal surveys where we used the theoretical framework and findings from the focus groups to investigate sense of belonging and community over time not only with LIS recruitment programs, but also compared to experiences in other institutions. We aim to now expand on the data collected primarily to complete interviews and disseminate findings. Interviews will provide nuanced data on how underrepresented students develop community within LIS recruitment programs, how this sense of community changes over time, which programmatic elements play a role in this evolution, how sense of community compares to experiences in other institutions, and how feelings in recruitment can scale to address isolation and gaps in support.

The iSchool Inclusion Institute (i3) is an undergraduate research and leadership development program that prepares students from underrepresented populations for graduate study and careers in the information sciences. Only 25 students from across the country are selected each year to become i3 Scholars. Those students undertake a yearlong experience that includes two summer institutes hosted by the University of Texas at Austin’s iSchool and a research project spanning the year. i3 prepares students for the rigors of graduate study and serves as a pipeline for i3 Scholars into internationally recognized information schools—the iSchools. Most importantly, i3 empowers students to create change and make an impact on the people around them.

In 2011, the National Science Foundation began requiring that all funded projects provide data management plans (DMPs) to ensure that project data, computer codes, and methodological procedures were available to other scientists for future use. However, the extent to which these data management requirements have resulted in more and better use of project data remains an open question. This project thus investigates the National Science Foundation's DMP mandate as a national science policy and examines the broad impacts of this policy across a strategic sample of five disciplines funded by the National Science Foundation. It considers the organization and structure of DMPs across fields, the institutions involved in data sharing, data preservation practices, the extent to which DMPs enable others to use secondary project data, and the kinds of data governance and preservation practices that ensure that data are sustained and accessible. Systematic investigation of the impact of DMPs and data sharing cultures across fields will assist funding agencies and research scientists working to produce reproducible and open science by identifying barriers to data archiving, sharing, and access. The principal investigators will use project findings to develop data governance guidelines for information professionals working with scientific data and to articulate best practices for scientific communities using DMPs for data management. This project aims to enhance understanding of the role data management plans (DMPs) play in shaping data life-cycles. It does so by examining DMPs across five fields funded by the National Science Foundation to understand data practices, archiving and access issues, the infrastructures that support data sharing and reuse, and the extent to which project data are later used by other researchers. In phase I, the investigators will gather a strategic sample of DMPs representing a wide range of data types and data retention practices from different scientific fields. Phase II consists of forensic data analysis of a subset of DMPs to discover what has become of project data. Phase III develops detailed case studies of research project data life-cycles and data afterlives with qualitative interviews and archival documentary analysis to help develop best practices for sustainable data preservation, access, and sharing. Phase IV will translate findings into data governance recommendations for stakeholders. The project thus contributes to research about contemporary studies of scientific data production and circulation while assessing the effect of DMPs as a national science policy initiative affecting data management practices in different scientific communities. The comparative research design and mixed methods enables theory building about cross-disciplinary data practices and data cultures across fields and advances knowledge within data studies, information management studies, and science and technology studies.

The information we generate on social media sites and in mobile device apps represents the fastest form of data creation and collection in the United States. However, these data traces are complicated to work with because they are varied, inter-dependent, and vulnerable to loss. In this Early Career Development project, Dr. Amelia Acker at the University of Texas at Austin, will conduct a three-year, qualitative investigation into the activities of engineers and designers at five institutions where social media software is being developed. This project to better understand developer cultures will aid archives, libraries, and museums as they develop and implement best practices for gathering and preserving social media collections.

In 2011, the National Science Foundation began requiring that all funded projects provide data management plans (DMPs) to ensure that project data, computer codes, and methodological procedures were available to other scientists for future use. However, the extent to which these data management requirements have resulted in more and better use of project data remains an open question. This project thus investigates the National Science Foundation's DMP mandate as a national science policy and examines the broad impacts of this policy across a strategic sample of five disciplines funded by the National Science Foundation. It considers the organization and structure of DMPs across fields, the institutions involved in data sharing, data preservation practices, the extent to which DMPs enable others to use secondary project data, and the kinds of data governance and preservation practices that ensure that data are sustained and accessible. Systematic investigation of the impact of DMPs and data sharing cultures across fields will assist funding agencies and research scientists working to produce reproducible and open science by identifying barriers to data archiving, sharing, and access. The principal investigators will use project findings to develop data governance guidelines for information professionals working with scientific data and to articulate best practices for scientific communities using DMPs for data management.     This project aims to enhance understanding of the role data management plans (DMPs) play in shaping data life-cycles. It does so by examining DMPs across five fields funded by the National Science Foundation to understand data practices, archiving and access issues, the infrastructures that support data sharing and reuse, and the extent to which project data are later used by other researchers. In phase I, the investigators will gather a strategic sample of DMPs representing a wide range of data types and data retention practices from different scientific fields. Phase II consists of forensic data analysis of a subset of DMPs to discover what has become of project data. Phase III develops detailed case studies of research project data life-cycles and data afterlives with qualitative interviews and archival documentary analysis to help develop best practices for sustainable data preservation, access, and sharing. Phase IV will translate findings into data governance recommendations for stakeholders. The project thus contributes to research about contemporary studies of scientific data production and circulation while assessing the effect of DMPs as a national science policy initiative affecting data management practices in different scientific communities. The comparative research design and mixed methods enables theory building about cross-disciplinary data practices and data cultures across fields and advances knowledge within data studies, information management studies, and science and technology studies.

The Texas Advanced Computing Center (TACC) at the University of Texas at Austin will acquire and deploy Stampede 2, a new, nearly 20 petaflop High Performance Computing (HPC) system. This system will be available to and accessed by thousands of researchers across the country. It will enable new computational and data-driven scientific and engineering, research and educational discoveries and advances. As a national resource, Stampede 2 will replace and surpass the current highly successful Stampede system. The new system will deliver over twice the overall performance as the current system in many dimensions most important to scientific computing, including computing capability, storage capacity, and network bandwidth. TACC and its academic partners will team with Dell, Inc. and Intel Corp. to procure and provide this system.  HPC is intrinsic to discovery across the science and engineering disciplines served by the NSF. This resource allows researchers to explore those scientific and engineer frontiers that require very large scale computations not otherwise possible. Over the life of Stampede 2, the system is expected to serve many thousands of researchers spanning all NSF-supported disciplines, as the current system has done. In addition to being an immediately productive resource for a large community of computational engineers and scientists, Stampede 2 will also continue the community on an evolutionary path to future "many core" computing technologies.  Stampede 2 will employ upcoming generations of Intel's Xeon and Xeon Phi processors, as well as the Intel Omni-Path network fabric. The system will maintain a familiar Linux-based software environment to insure a smooth migration of the large existing user base to the new system. The system and its software stack will be designed to support traditional large scale simulation users, users performing data intensive computations, as well as emerging classes of new and non-traditional users to high performance computing. Stampede 2 will support breakthrough discoveries and advances across a wide range of research topics.

While the broad goals of socially responsible artificial intelligence (AI) appear clear in the abstract, how can we translate such goals into practice for a real problem facing our society today? We consider the following challenge: How can we design responsible AI technologies to curb the digital spread of misinformation?  Exploring real use cases and interface designs, we develop prototype AI applications and user-centered evaluations to remedy situations in which misinformation circulates online.

Artificial intelligence (AI) systems are increasingly used to assist humans in making high-stakes decisions, such as online information curation, resume screening, mortgage lending, police surveillance, public resource allocation, and pretrial detention. While the hope is that the use of algorithms will improve societal outcomes and economic efficiency, concerns have been raised that algorithmic systems might inherit human biases from historical data, perpetuate discrimination against already vulnerable populations, and generally fail to embody a given community's important values. Recent work on algorithmic fairness has characterized the manner in which unfairness can arise at different steps along the development pipeline, produced dozens of quantitative notions of fairness, and provided methods for enforcing these notions. However, there is a significant gap between the over-simplified algorithmic objectives and the complications of real-world decision-making contexts. This project aims to close the gap by explicitly accounting for the context-specific fairness principles of actual stakeholders, their acceptable fairness-utility trade-offs, and the cognitive strengths and limitations of human decision-makers throughout the development and deployment of the algorithmic system.  To meet these goals, this project enables close human-algorithm collaborations that combine innovative machine learning methods with approaches from human-computer interaction (HCI) for eliciting feedback and preferences from human experts and stakeholders. There are three main research activities that naturally correspond to three stages of a human-in-the-loop AI system. First, the project will develop novel fairness elicitation mechanisms that will allow stakeholders to effectively express their perceptions on fairness. To go beyond the traditional approach of statistical group fairness, the investigators will formulate new fairness measures for individual fairness based on elicited feedback. Secondly, the project will develop algorithms and mechanisms to manage the trade-offs between the new fairness measures developed in the first step, and multiple existing fairness and accuracy measures. Finally, the project will develop algorithms to detect and mitigate human operators' biases, and methods that rely on human feedback to correct and de-bias existing models during the deployment of the AI system.

The gig economy is characterized by short-term contract work performed by independent workers who are paid in return for the "gigs" they perform. Example gig platforms include Uber, Lyft, Postmates, Instacart, UpWork, and TaskRabbit. Gig economy platforms bring about more job opportunities, lower barriers to entry, and improve worker flexibility. However, growing evidence suggests that worker wellbeing and systematic biases on the gig economy platforms have become significant societal problems. For example most gig workers lack financial stability, have low earning efficiency and lack autonomy, and many have health issues due to long work hours and limited flexibility. Additionally, gig economy platforms perpetuate biases against already vulnerable populations in society. To address these problems, this project aims to build a community-centered, meta-platform to provide decision support and data sharing for gig workers and policymakers, in order to develop a more vibrant, healthy, and equitable gig economy. The project involves three major research activities. (1) Working with gig workers and local policymakers to understand their concerns, challenges, and considerations related to gig worker wellbeing, as well as the current practices, problems, and biases of existing gig economy platforms. (2) Developing a data-driven and human-centered decision-assistance environment to help gig workers make "smart" decisions in navigating and selecting gigs,and provide a macrolevel perspective for policymakers working to balance their diverse set of objectives and constraints. (3) Deploying and evaluating whether and how the above environment addresses the fundamental problems of worker wellbeing and systematic biases in the gig economy.

The University of Tennessee at Knoxville; The University of Illinois, Urbana-Champaign; and the University of Texas, Austin are collaborating on the IDEA (Innovation, Disruption, Enquiry, Access) Institute on Artificial Intelligence (AI). This institute will address a gap in education and training for AI leaders in the library and information field through a one-week intensive, interactive, evidence-based, and applications-oriented professional development program for library and information professionals. The Institute will create two cohorts of leaders in knowledge and skills in AI to evaluate and implement in library and information environments. The curriculum will incorporate conceptual, technical, social, and applied aspects, including ethical issues of AI. The project will have national impact by sparking future innovation, collaboration, and dissemination of AI in library and information environments. It is supported by the ALA Center for the Future of Libraries and sustained through the Association of Information Science and Technology.