Date: November 11, 2025 (Tuesday)
Time: 14:30 – 17:30
Venue: HW612B, 6/F Haking Wong Building, HKU
Programme:
14:30 - 14:35 Opening speech & introduction by Prof WY Szeto
14:35 - 15:20 "Urban Mobility, Air Pollution, and Environmental Justice"
by Professor Junshi XU
15:20 - 16:05 "Representative Agents for Traffic Modeling: From Fully LLM-Driven to LLM
Guided Learning" by Professor Jiayang LI
16:05 - 16:20 Break
16:20 - 17:30 Brainstorming and discussion on research collaborations /
ITS initiatives (led by Professor WY Szeto)
All categories of ITS fellows are welcome. The abstracts of the talks and the bios of Prof. Xu and Prof. Li can be checked below.
Urban Mobility, Air Pollution, and Environmental Justice
Professor Junshi XU
Department of Geography, The University of Hong Kong
Abstract:
Urban mobility is key to building equitable and sustainable cities, especially for populations facing socioeconomic disadvantage, transport poverty, and high exposure to traffic-related air pollution (TRAP). This seminar introduces the PEACE framework (People oriented, Equitable, Accessible, Community focused, and Environmental), which integrates engineering, data science, and social perspectives to address urban mobility and environmental justice challenges. Drawing on Canadian research that applied environmental sensing, machine learning, and computer vision to examine TRAP and its unequal distribution, the seminar also reflects on community engagement and institutional initiatives in Toronto that revealed the challenges faced by disadvantaged groups and identified pathways for local interventions. New insights from Hong Kong extend these approaches through high-resolution sensing and vehicle-class-resolved modelling, showing how goods vehicles and buses contribute to disparities in exposure, particularly among low-income and public housing residents. The seminar demonstrates how integrated methods can inform targeted emission mitigation and equitable transport policies, advancing more inclusive, resilient, and sustainable cities.
Bio:
Dr. Junshi Xu is an Assistant Professor in the Department of Geography at The University of Hong Kong, where he also serves as Deputy Program Director of the Master of Transportation Policy and Planning (MTPP). He earned his Ph.D. in Transportation Engineering from the University of Toronto in 2020.
His research integrates environmental sensing, machine learning, and computer vision to investigate transport-related air pollution and environmental justice. He has led and contributed to projects in Canada and Hong Kong that combine high-resolution mobile monitoring, vehicle-class-resolved modelling, and community engagement to uncover disparities in exposure, particularly among disadvantaged groups. These studies aim to inform equitable policies and advance sustainable urban mobility solutions.
Dr. Xu has published over 45 peer-reviewed articles in journals such as Environmental Science & Technology, Environmental Pollution, Science of the Total Environment, and Transportation Research Part D. He served on the Transportation Research Board’s Committee on Air Quality and Greenhouse Gas Mitigation, is Executive Guest Editor for Transportation Research Part D, and has been Guest Editor for the Journal of Atmosphere and the Journal of Computational Urban Science.
Representative Agents for Traffic Modeling:
From Fully LLM-Driven to LLM-Guided Learning
Professor Jiayang LI
Department of Data and Systems Engineering, The University of Hong Kong
Abstract:
Large language models (LLMs) are increasingly used as behavioral proxies for self-interested travelers in agent-based traffic models. Although more flexible and generalizable than conventional models, the practical use of these approaches remains limited by scalability due to the cost of calling one LLM for every traveler. Moreover, it has been found that LLM agents often produce unstable day-to-day dynamics. To address these challenges, we propose to model each homogeneous traveler group facing the same decision context with a single representative LLM agent who behaves like the population's average, maintaining and updating a mixed strategy over routes that coincides with the group's aggregate flow proportions. Each day, the LLM reviews the travel experience and flags routes with positive reinforcement that they hope to use more often, and an interpretable update rule then converts this judgment into strategy adjustments using a tunable (progressively decaying) step size. The representative-agent design improves scalability, while the separation of reasoning stabilizes learning. In classic traffic assignment settings, we find that the proposed approach converges rapidly to the user equilibrium. In richer settings with income heterogeneity, multi-criteria costs, and multi-modal choices, the generated dynamics remain stable and interpretable. Moreover, it reproduces behavioral patterns well-documented in psychology and economics, for instance, the decoy effect in toll and non-toll road selection and the greater willingness-to-pay for convenience among higher-income travelers.
Bio:
Prof. Jiayang Li is an Assistant Professor in the Department of Data and Systems Engineering and a Fellow of the Institute of Transport Studies at The University of Hong Kong. Prof. Li earned his Ph.D. in transportation systems analysis from Northwestern University in 2024. Before that, he obtained his B.S. in mathematics from Tsinghua University in 2019. Prof. Li's research aims to bring together optimization, game theory, and machine learning to address operations research challenges, particularly within transportation and mobility systems. His work has been published across diverse fields, including leading transportation journals (Transportation Science and Transportation Research Part B) as well as top machine learning conferences (NeurIPS and ICML).
