Hang Su
Papers
1
Total Citations
14
H-Index
1
About
Hang Su is an emerging researcher at the forefront of AI safety and embodied intelligence, with a particular focus on the robustness and security of large language model (LLM)-based systems. His most recognized work investigates the vulnerabilities of decision-level processes in LLM-driven embodied agents, exploring how adversarial attacks can compromise the perception and planning capabilities of AI systems operating in real-world environments. This research addresses a critical gap in understanding how intelligent agents — designed to interpret complex language instructions and execute sophisticated multi-step tasks — can be manipulated or deceived, raising important implications for the safe deployment of autonomous systems. With 14 citations on his leading paper published in 2024, Su's work has already begun attracting meaningful attention within the AI security and robotics communities, a notable achievement given the recency of the publication. His research sits at a compelling intersection of natural language processing, embodied AI, and adversarial machine learning — fields of rapidly growing importance as AI systems are increasingly deployed in physical and high-stakes environments. For students and researchers interested in AI robustness and trustworthy autonomy, Su's contributions offer valuable and timely insights.
Research Focus
Key Achievements
Top Papers
- 1