Warren He
Papers
1
Total Citations
52
H-Index
1
About
Warren He is a prominent researcher in the field of artificial intelligence security, with a particular focus on adversarial machine learning and the robustness of deep learning systems. His most cited work, "Characterizing Attacks on Deep Reinforcement Learning" (2019, 52 citations), provides a foundational analysis of vulnerabilities in DRL models, demonstrating how small perturbations to observations can compromise model performance. He critically evaluates existing attack methodologies, highlighting their impractical assumptions—such as full access to victim models or excessive computational demands—and paving the way for more realistic security assessments. Beyond this, He’s contributions extend to developing defenses and understanding the broader implications of adversarial threats in AI. His research has been instrumental in shaping how the community approaches the safety and reliability of autonomous systems, from robotics to game-playing agents. With a strong citation impact and a focus on bridging theoretical attacks with practical constraints, Warren He is recognized as a key voice in the ongoing effort to build trustworthy AI. His work continues to inspire students and researchers aiming to secure next-generation intelligent systems.
Research Focus
Key Achievements
Top Papers
- 1Characterizing Attacks on Deep Reinforcement Learning52 citations · 2019