Biography
Senkang "Forest" Hu (胡森康, feel free to call him "Forest") is currently a Ph.D. candidate at the Department of Computer Science, City University of Hong Kong, under the supervision of Fellow of ACM & IEEE & AAAS, Chair Prof. Yuguang "Michael" Fang. Previously, he was a Visiting PhD at the Department of Computer Science, University of Oxford, working with Prof. Jiarui Gan. He received his BSc degree from the School of Information and Electronics, Beijing Institute of Technology in 2022.
His current research focuses on long-horizon LLM agents, agentic reinforcement learning, and self-evolving agents. His long-term goal is to build agents that solve complex tasks, learn from interaction, and continually improve with limited external supervision. His work has appeared at top-tier conferences, including ICML, NeurIPS, CVPR, AAAI (Oral), and ICRA, and in top-tier journals, including TMC, TITS, and ToN. He also serves as a reviewer or technical program committee member for ICML, ICLR, NeurIPS, ICRA, TMC, ToN, TITS, etc. He received the CYCSGS Scholarship, CityUHK's highest honor for PhD students, as the sole awardee in the Dept. of CS. He also received the Goertek Scholarship as one of only two awardees in the Dept. of CS. He received the Best Paper Award at GLOBECOM'25.
Education & Experience
Dept. of Computer Science, conducting research on agentic reinforcement learning, self-evolving agents, and game theory.
Working with Prof. Jiarui Gan
Conduct research on Agentic Reasoning and Reinforcement Learning, proposed InfoReasoner (ICML'26).
Working with Prof. Yuzhi Zhao
Dept. of Computer Science, supervised by Chair Prof. Yuguang "Michael" Fang, ACM/IEEE/AAAS Fellow
Chow Yei Ching School of Graduate Studies Scholarship (Top 0.3%, sole awardee in CS Dept.)
Electronic Information Engineering
National Scholarship (Top 1%)
Research Vision
Building LLM agents that solve long-horizon tasks, learn through reinforcement, and continually self-evolve.
Long-Horizon LLM Agents
Developing LLM agents that maintain coherent plans, manage evolving task states, use tools, and recover from errors across extended interaction trajectories.
- Planning & Multi-Step Reasoning
- Search, Retrieval & Tool Use
- Efficient Long-Horizon Inference
Agentic Reinforcement Learning
Designing reinforcement learning and credit-assignment methods that optimize multi-step agent trajectories from sparse, delayed, and outcome-based feedback.
- Reward Design & Learning Signals
- Long-Horizon Credit Assignment
- Exploration & Policy Optimization
Self-Evolving Agents
Enabling agents to learn from experience, acquire reusable skills, and autonomously improve their capabilities with limited external supervision.
- Experience-Driven Self-Improvement
- Context & Memory Evolution
- Skill Acquisition & Transfer
Publications
Honors & Awards
Research Activity Fund
City University of Hong Kong
Best Paper Award
IEEE Communications Society
UAV-enabled Computing Power Networks: Task
Completion Probability Analysis at IEEE Global Communications Conference
(GLOBECOM'25)
IEEE ComSoc Student Grant
IEEE Communications Society
Conference Travel Grant for Research Degree Students
City University of Hong Kong
Chow Yei Ching School of Graduate Studies Scholarship, Top 0.3%
City University of Hong Kong
10 awardees university-wide, and the sole awardee in the Department of Computer Science.
Research Tuition Scholarship
City University of Hong Kong
IEEE ComSoc Student Travel Grant
IEEE Communications Society
Institutional Research Tuition Grant
City University of Hong Kong
HK$170,000 to cover part of tuition fees during PhD.
InnoHK Award Funding
The Government of the Hong Kong Special Administrative Region
Around HK$900,000 to support research and PhD career.
National Scholarship
Ministry of Education of the People's Republic of China
The highest national honors for undergraduate students in China.
Example of Scientific Research Innovation, Top 1%
Beijing Institute of Technology
First Class Academic Scholarship, 4 times
Beijing Institute of Technology
Excellent Student of Beijing Institute of Technology, 3 times
Beijing Institute of Technology
Services
Conference Technical Program Committee & Reviewer
- ICML
- ICLR
- NeurIPS
- AAAI
- CVPR
- ICCV
- CSCW
- ACM MM
- ICRA
- IROS
- ECCV
- ICCC
Journal Reviewer
- ACM Computing Surveys
- IEEE TMC
- IEEE/ACM ToN
- IEEE TITS
- IET Information Security
- IEEE TNSM
- ACM TAAS
- Information Fusion
- Pattern Recognition
- IEEE RA-L
- Proceedings of the IEEE
- IEEE TVT
- IEEE TPDS
Teaching Assistant
- Cloud Computing, 2026 Spring
- Cloud Computing: Theory and Practice, 2026 Spring
- Info Security for eCommerce, 2025 Fall
- Cloud Computing, 2025 Spring
- Computer Networks, 2024 Fall
- Cloud Computing: Theory and Practice, 2024 Spring