Home Stakeholders Research & Innovation Stakeholders Data Institute for Societal Challenges (DISC) advancing data-enabled research

Data Institute for Societal Challenges (DISC) advancing data-enabled research

The Data Institute for Societal Challenges (DISC) is setting a new benchmark for cutting-edge advances in artificial intelligence, machine learning, and real-world applications driven by advancements in data-enabled research

Data science is becoming increasingly critical to current and future discovery and innovation in the state of Oklahoma, the nation, and the world. The University of Oklahoma (OU) is redefining the landscape as a leading center of excellence in data science research and data-driven solutions.

With the introduction of the Data Institute for Societal Challenges (DISC), OU is swiftly advancing the forefront of discovery through its investment in highly skilled researchers, top-tier research facilities, and partnerships that bridge the academic, private, industrial, and governmental sectors. DISC is setting a new benchmark for cutting-edge advances in artificial intelligence, machine learning, and real-world applications driven by advancements in data-enabled research. OU’s research investment in these technologies will profoundly impact society, from breakthroughs in the development of robust and predictive software and guidance systems used by the U.S. Air Force, Army, and Department of Homeland Security, to advances in precision medicine that aid in the early detection and more effective treatment of disease, to the development of more accurate, timely, physics-informed weather and severe storm predictions and forecasts, all of which protect millions of lives.

The development of ecologically sustainable communities and energy grids, as well as transformed modern supply chains around the world, are not just goals; they are achievable through the data science research endeavours at OU.

Human-guided artificial intelligence and machine learning

Artificial intelligence (AI) and machine learning (ML) are unlocking the next generation of advances in science, engineering, and other disciplines. AI and ML are currently used in a variety of capacities, such as developing pricing models for utilities and assisting with drug and protein discovery. The goal of developing robust, trustable, explainable, and fair AI and ML is to help users understand how these systems produce results, remove biases in how these systems work, and promote the further use and adoption of AI and ML. Developing interactive, human-guided techniques that harness user expertise and knowledge has the potential to improve the robustness and performance of purely automated AI/ML techniques and to increase understandability and trust in the results. DISC is building on OU’s strength in integrating AI/ML specialists with social scientists, cognitive psychologists, political scientists, biologists, and engineers to create cross-disciplinary AI/ML solutions for advancing theories to solve societal challenges.

OU is currently using and developing innovative AI and ML convergent research that brings together expertise from science, engineering, and social science, as well as computer and data science. Specifically, OU is home to one of five NSF Artificial Intelligence Institutes that is creating trustworthy AI methods for environmental scientists while revolutionizing our understanding of atmospheric phenomena. DISC is leading an NSF Planning Grant to create a roadmap for ensuring sustainable agricultural production and communities using AI and ML.

Human-computer teaming

Human-computer teaming, a critical problem space for data-enabled research, is the efficient and effective integration of humans and complex machines. Effectively blending human and machine capabilities while accounting for the unique strengths and limitations of both will enable us to address complex problems, such as introducing autonomous vehicles to roadways, disaster recovery, and medical diagnostics. Over the last 25 years, discoveries in cognitive neuroscience and technological advancements in ML have led to new insights into the underlying capacities needed to support effective human-computer teams and overcome the limited contextual knowledge, cognitive inflexibility, and opaqueness of AI and ML. Transforming intelligent machines from tools to teammates requires cognitive and computational models of beliefs, desires, intentionality, and capabilities. Our approach draws on the expertise of cognitive psychologists, device designers, human factors engineers, decision-making and risk-perception researchers, user experience researchers, and computer scientists. Each supplies unique insights into how humans work in team science and process design.

Predictive analytics

Predictive analytics, the practice of analyzing and mining data and historical trends to make a prediction or find a solution, is empowering research, policies, and decisions that affect our daily lives. Predictive analytics is needed to solve some of the world’s most pressing problems. OU researchers have expertise in creating tools and systems capable of extracting, assimilating, and analyzing data for accurate, timely, reliable forecasts and predictions with quantifiable uncertainty. Predictive analytics are already impacting our daily lives by transforming drug discovery, vaccine production and enabling effective, personalized treatments and new pathways for the future of health. Furthermore, predictive analytics has led to earlier and more precise forecasts for the impact of various events. However, research is needed to identify and correct for some potential pitfalls of predictive analytics.

Our goal is to develop analytics in such a way that we can be constantly correcting for unintended consequences or adverse implications.

Scalable, high-performance software and hardware architectures

Cutting-edge AI and ML algorithms are computationally intensive and must process large quantities of data to perform well. The slow training and performance of AI and ML algorithms are attributable to large volumes of data necessary for learning and their increasing computational complexity. This challenge limits their wider use in real-time applications. Interdisciplinary teams researching solutions to global grand challenges need scalable and elastic solutions powered by emerging computing architectures, cloud-based storage, and the processing of globally distributed data.

At OU, we have research teams developing and applying these new approaches to power data-enabled science as they work to solve problems such as early detection and response to emerging infectious diseases, sub-surface carbon sequestration to create net-zero carbon energy and sustainable environmental solutions, improved maintenance and life-extension of critical defense aircraft; improved medical treatments and strategies to reduce health disparity; digital preservation of cultural artefacts and understanding of ancient peoples and societies; and social justice and reduced disparity among communities. The ongoing, synergistic research at OU AI and ML seeks to solve today’s challenges using tomorrow’s technologies.

Learn more about these efforts and the Data Institute for Societal Challenges at the University of Oklahoma at ou.edu/disc.