Zihe Sun
Papers
1
Total Citations
2
H-Index
1
About
Zihe Sun is a researcher at the forefront of integrating formal methods with deep reinforcement learning, specializing in the synthesis of safe and verifiable control policies for autonomous systems. Their work addresses the critical challenge of enabling agents with unknown stochastic dynamics to reliably satisfy complex temporal logic specifications, particularly those expressed in Linear Temporal Logic (LTL). Sun’s key contributions include the development of mission-driven exploration strategies that dramatically accelerate the learning of policies maximizing satisfaction probabilities, bridging the gap between high-level task specifications and low-level control. Their 2023 paper, "Mission-driven Exploration for Accelerated Deep Reinforcement Learning with Temporal Logic Task Specifications," has already garnered attention for its novel approach to guiding exploration in sparse-reward environments. By combining formal verification with data-driven methods, Sun is helping to build a foundation for trustworthy AI systems that can guarantee performance in safety-critical applications such as robotics and autonomous navigation. Their work stands at the intersection of control theory, machine learning, and formal logic, promising to make autonomous agents both more capable and more reliable.
Research Focus
Key Achievements
Top Papers
- 1