IDIA Predoctoral Fellows

Jinyi Kim

Project Title: A Multi-level Approach to Measuring Risks for Positive Youth Outcomes

Youth involved with the juvenile justice system tend to show higher rates of adverse childhood experiences, increased health needs, and learning disabilities. Many of these already vulnerable youth have received interventions through other systems (e.g., schools, mental health), which have been ineffective. We move beyond conventional approaches to studying youth risk, which are limited by (1) a narrow focus on juvenile-justice samples and delinquency outcomes and (2) a lack of attention to community-level factors. The present study will increase the use of youth-oriented research by moving beyond questions of “what works” to developing knowledge of “how to use” existing knowledge. Specifically, the proposed research will extend the Positive Youth Development (PYD) Framework by theorizing and developing a standardized measure of community-level risk factors (e.g., service deserts) through computational approaches. The proposed framework and measurements will inform youth-related policymakers and practitioners in designing and assessing multi-level interventions to promote youth’s development.

Team of faculty mentors:

Myeong Lee, Assistant Professor, Department of Information Sciences and Technology, School of Computing, College of Engineering and Computing

JoAnn Lee, Associate Professor, Department of Social Work, College of Health and Human Services

Hamdi Kavak, Assistant Professor, Department of Computational and Data Sciences, College of Science

Learn more about Jinyi's research:

Soelem Aafnan Bhuiyan

Project Title: Developing a Data Assimilation Technique to Improve the Monitoring and Forecasting of Coastal Hazards 

Hurricanes, cyclones, and typhoons are major global natural hazards causing frequent and recurrent devastation worldwide. Providing accurate and reliable water levels and flood inundation extents to inform early warning systems remains a challenge at the operational level. The NASA Surface Water and Ocean Topography (SWOT) mission (planned to be launched in November 2022) and its observations of sea water levels have the potential to fundamentally transform our ability to predict flooding from tropical storms. 

This IDIA pre-doctoral fellowship will allow me to test the hypothesis that, by assimilating SWOT products within a storm surge model, we can improve the estimation of storm surge process dynamics and operational forecasting. This project will advance our understanding of such processes to build resilience of our coastal infrastructure and economy, therefore, promoting security, prosperity and the well-being of coastal populations. 

Team of faculty mentors:

Viviana Maggioni, Associate Professor, Sid and Reva Dewberry Department of Civil, Environmental, and Infrastructure Engineering, School of Engineering, College of Engineering and Computing

Celso Ferreira, Associate Professor, Sid and Reva Dewberry Department of Civil, Environmental, and Infrastructure Engineering, School of Engineering, College of Engineering and Computing

Learn more about Soelem’s research:

Facundo Nicolas Airaudo

Project Title: Optimal HVAC Outlay to Minimize Pathogen Transmission

The experiments and simulations carried out before and during the COVID-19 pandemic have made it clear that current Heating, Ventilation, and Air Conditioning (HVAC) systems can lead to very unfavorable airflows, resulting in increased pathogen transmission. This project aims to minimize the amount of viral matter inhaled by human beings in built environments. This will be carried out by modeling and solving optimization problems within the physical constraints of the equations describing fluid dynamics, temperature, and particle transport. The control parameters are viral matter source, location of vents, inflow velocities, etc. The project can potentially help update the CDC guidelines on how to keep people safe in built environments, not just from COVID but other airborne pathogens as well. It will also help provide optimal outlays for HVAC vents with minimal pathogen transmission. The ideas generated under this project are equally applicable to airplanes, hospitals, stores, courtrooms, classrooms, etc.

Team of faculty mentors:

Rainald Lohner, Professor, Department of Physics and Astronomy, College of Science 

Harbir Antil, Professor, Department of Mathematical Sciences, College of Science 

Juan Raul Cebral, Professor, Department of Bioengineering, School of Engineering, College of Engineering and Computing

Learn more about Facundo’s research:

Michael Timothy Crawshaw

Project Title: Multiple Guarantees in Distributed Deep Learning: Computation, Communication, and Sample Efficient Algorithms

The goal of our research is to develop distributed deep learning algorithms with theoretical guarantees and deploy these algorithms for diverse applications. Deep learning is a powerful area of machine learning with two barriers for real-world deployment: high resource demand and lack of theoretical guarantees. We aim to address these weaknesses by developing the theory and practice of distributed deep learning, that is, deep learning methods which utilize computational power and data from many communicating devices. Distributed learning can ease the resource requirements of deep learning by eliminating the need for an expensive, monolithic training environment, lowering the barrier to entry. Also, by developing distributed learning algorithms with provable guarantees of computation, communication, and sample efficiency, we aim to enable confident real-world deployment of deep learning models. Our work will enable easier access to deep learning while increasing its real-world scope and energy demands, with the ultimate goal of improving societal well-being.

Team of faculty mentors:

Mingrui Liu, Assistant Professor, Department of Computer Science, School of Computing, College of Engineering and Computing

Wanli Qiao, Assistant Professor, Department of Statistics, School of Computing, College of Engineering and Computing

Ozlem Uzuner, Associate Professor, Department of Information Sciences and Technology, School of Computing, College of Engineering and Computing

Learn more about Michael's research: