Ling Xiong
Papers
3
Total Citations
10
H-Index
2
About
Ling Xiong is an emerging researcher specializing in the robustness and security of deep reinforcement learning (DRL), with a particular focus on adversarial attacks and their implications for real-world robotic control systems. His work addresses a critical challenge in applied AI: the vulnerability of DRL agents to environmental perturbations and observation inaccuracies that arise when transitioning from simulated training environments to physical deployment. Xiong's most notable contribution, "Rethinking Adversarial Attacks in Reinforcement Learning from Policy Distribution Perspective" (2025, 5 citations), introduces a fresh conceptual framework that moves beyond targeting individual sampled actions, offering a more principled evaluation of agent robustness. Complementing this, his research on adaptive gradient-masked adversarial attacks and state-aware perturbation optimization advances the design of white-box attack strategies tailored specifically for reinforcement learning contexts, rather than borrowing inadequately from supervised learning paradigms. Collectively accumulating 10 citations across three publications released in 2025 alone, Xiong demonstrates a rapidly growing scholarly presence. His research holds significant implications for the safe deployment of autonomous robots in uncontrolled environments, making his work increasingly relevant to both the academic community and practitioners developing robust AI-driven robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3