Abstract: Reinforcement learning (RL) has demonstrated remarkable success in domains such as gaming, robotics, and large language models (LLMs), sparking growing interest in its potential for real-world decision-making. However, applying RL to such settings introduces significant challenges, particularly related to data sparsity, model generalization, and real-world constraints. In this talk, I will examine how these challenges affect the practicality of RL in urban domains and present our recent efforts to address them. Specifically, I will share our preliminary efforts for bridging the simulation-to-reality gap through uncertainty quantification, integrating language models for policy abstraction, and developing testbeds for empirical validation. These directions aim to move current data analytics from descriptive and predictive insights to truly actionable intelligence.

Bio: Hua Wei is an Assistant Professor in the School of Computing and Augmented Intelligence (SCAI) at Arizona State University (ASU). His research interests lie in data mining and machine learning, with a particular emphasis on spatio-temporal data mining and reinforcement learning. His work has received multiple Best Paper Awards, including honors from ECML-PKDD and ICCPS. His research has been published in leading conferences and journals in machine learning, artificial intelligence, data mining, and control, such as NeurIPS, ICML, AAAI, CVPR, KDD, IJCAI, ECML-PKDD, and WWW. His research has been supported by various funding  agencies, including the National Science Foundation (NSF), the Department of Energy (DoE), and the Department of Transportation (DoT). Dr. Wei is a recipient of the NSF CAREER Award in 2025, Amazon Research Award and Cisco Research Award.