Xujie Song

Papers

Total Citations

H-Index

About

Xujie Song is a rising researcher at the forefront of reinforcement learning (RL), with a focused interest in enhancing the robustness and real-world applicability of RL-based control systems. His most-cited work, "Smooth Filtering Neural Network for Reinforcement Learning" (2024), addresses a critical bottleneck in the field: the tendency of learned control policies to produce jerky, non-smooth actions when faced with sensor noise or environmental disturbances. By integrating a smooth filtering mechanism into the neural network architecture, Song's approach significantly improves the stability and reliability of RL agents in complex tasks such as vehicle tracking control and obstacle avoidance. Though early in his career, his contributions are already gaining traction, with his flagship paper accumulating 3 citations and signaling a growing interest in bridging the gap between theoretical RL and practical, noise-tolerant deployment. Song’s work is particularly notable for its potential impact on autonomous systems, where smooth, safe decision-making is paramount. As he continues to develop more robust learning frameworks, Xujie Song is poised to become a key voice in the next generation of RL research.