Jundong Li joins research effort funded by National Science Foundation's Strengthening American Infrastructure Program.The water well has a mythical and storied place as the heart of a community. Our ancestors approached underground springs as a providential place, for which they gave thanks in the form of small tokens — the original wishing well.
To protect their source of fresh water, settlers and townsfolk built protective structures such as well houses, which became a community gathering place. At a time when settlements dotted the landscape, the town well provided an opportunity to generate good will through hospitality and to learn about events taking place beyond the town's borders.
As cities grew and technology advanced, industrial systems accessed and moved freshwater from distant sources. Yet, the U.S. Environmental Protection Agency reports that well water remains the primary source of drinking water for more than 23 million U.S. households. These wells continue to serve the needs of diverse communities, especially when public water systems fail due to floods or other natural disasters.
Jundong Li, assistant professor of electrical and computer engineering, computer science and data science, has joined a new research project to help ensure the water well continues to support community vitality, resilience and quality of life.
The project, “Dynamic Coupling of Physical and Social Infrastructures: Evaluating the Impact of Social Capital on Access to Safe Well Water,” is led by Associate Professor Ryan Qi Wang and Assistant Professor Kelsey Pieper, both with the Department of Civil and Environmental Engineering at Northeastern University. A $750,000 grant from the National Science Foundation Strengthening America's Infrastructure Program supports their research.
Wang and Pieper will explore the use of cell phone-based mobility data to visualize and describe the social networks that people form, and how they generate and use social capital to maintain private wells before and when disaster strikes. The resulting social network analysis can help public safety officials improve their predictions of well water quality in their post-storm damage assessments, and aid extension services and health departments in disaster planning.
Li will develop a fairness-aware AI framework to address possible biases in both the data and the deep learning models that generate the social network analyses. Li's research will compensate for data scarcity in underserved communities, a known weakness of cell phone mobility data.
“We care about the interactions and behaviors of people from different areas,” Li said. “I'm trying to develop fairness tools to mitigate this bias. My framework will help the team accurately quantify the social capital that circulates and accrues in underrepresented communities, to ensure that people from different communities and areas have equitable access to water during disasters.”
Fairness-aware AI is a focus of Li's team and a research strength of the Charles L. Brown Department of Electrical and Computer Engineering.