Opening
I am always actively looking for self-motivated students to join my research group as PhD students, research assistants (long/short-term/remote), and interns. Please first fill out the form and then send your inquiry with the subject Your Name + Position (PhD/RA/Intern) + Start Date to taoli.cityuhk@gmail.com, with your CV and transcripts attached.
Reinforcement Learning in Partially Observable Stochastic Games
All Year RoundPreferred Qualifications: POMDP, Policy Optimization, Dynamic Programming in Markov (Stochastic) Games
Conventional RL algorithms either assume perfect state observation, common information or centralized training to avoid belief hierarchies and nonstationarity among agents under partial observability. This research thrust aims to establish a new RL paradigm that addresses each individual agent's rationa learning under information asymmetry in games.
LLM Agent for Autonomous Cybersecurity
All Year RoundPreferred Qualifications: LLM Fine-tuning, RL Implementation, Network Security Basics
RL-based Cyber Agent, while displaying a decent level of autonomy in making security decisions, relies heavily on stylized mathematical modeling of network environments, which is often oversimplied and sometimes even detached from reality. This research thrust aims to build an LLM-ehanced agentic AI framework for autonomous security decision-making in computer and communication networks with minimal human intervention and decent transparency, explanability, and accountability.