Kecen Li

About

Kecen Li is a rising researcher at the intersection of artificial intelligence and cybersecurity, with a primary focus on reinforcement learning (RL) and its vulnerabilities. His most notable contribution is the development of **Baffle**, a novel framework that reveals how backdoor attacks can be stealthily hidden within offline RL datasets. This work, published in 2024 and already garnering 5 citations, exposes a critical security gap in offline RL—a paradigm prized for its ability to learn from pre-collected data without costly environment interactions. By demonstrating how malicious data providers can embed hidden triggers that compromise learned policies, Li’s research serves as an essential wake-up call for the safe deployment of RL in real-world applications like robotics and autonomous systems. His work not only advances the understanding of adversarial machine learning but also provides a foundation for developing more robust defense mechanisms. As a young scholar, Kecen Li is quickly establishing himself as a key voice in trustworthy AI, bridging the gap between reinforcement learning theory and practical security concerns.

Research Focus

Key Achievements

Top Papers

1. 1
   Baffle: Hiding Backdoors in Offline Reinforcement Learning Datasets

   5 citations · 2024

Key Collaborators

Contact & Links