Binghamton's Yingxue Zhang Wins $584,649 NSF CAREER Award for Smart-City AI

2 articles · Updated · Newswise · Aug 10

$584,649 will fund Yingxue Zhang’s five-year NSF CAREER project to build AI models that improve urban decision-making in areas such as traffic, transit and mobility.
Zhang’s research centers on offline reinforcement learning, which learns from existing GPS, traffic and public-transit data instead of trial-and-error methods she says would be too risky in real city settings.
The project aims to tackle messy spatial-temporal data, limited information from each individual and “distribution shift” between collected data and changing real-world conditions, with testing planned in simulated environments.
University of Maryland, University of Pittsburgh and Hong Kong partners will help trial the policies, while Zhang also plans new coursework, K-12 outreach and industry workshops tied to the research.
By the end of the grant, Zhang said the resulting models and data will be open source, with the goal of supporting broader smart-city research and AI workforce development.