TIME & LOCATION
Dec 4, 2025 9:30-11:30
Room 322, School of Transportation
INTRODUCTION
As the digital transformation of transportation accelerates, AI agents are reshaping traffic optimization, route planning, and mobility services. This talk introduces the key concepts and hands-on skills needed to build AI agents with modern frameworks. We will dissect the core cognitive loop that powers intelligent systems—“Think-Plan-Act-Observe”—showing how it scales from simple conversational bots to sophisticated decision-making engines capable of real-time traffic analysis and autonomous fleet dispatch.
ABOUT THE LECTURER
Ma Jiangshan is Algorithm Director at Trip.com Group. He received his B.S. (2007) and M.S. (2009) from Southeast University and his Ph.D. (2012) from Tokyo Institute of Technology. His research spans multi-task recommender systems, large-model agents, causal machine learning, GIS, routing algorithms, and multi-agent traffic simulation. In 2014 he led the development of Asia’s first multi-agent traffic-simulation system, earning Japan’s Ministry of Land, Infrastructure and Transport Award for Advanced Technology. His 2018 real-time dynamic-carpooling platform was featured on CCTV and now serves dozens of Chinese cities. The 2021 multimodal itinerary engine he spearheaded creates ticket-free travel options during peak periods, influencing more than 3 billion passenger trips annually and winning the First-Class Science & Technology Progress Award from the China Communications and Transportation Association.