Weihang Yuan
Papers
1
Total Citations
7
H-Index
1
About
Weihang Yuan is a researcher advancing the intersection of deep reinforcement learning and active perception, with a focus on enabling intelligent agents to make decisions under partial observability. His most-cited work, "A Layered Architecture for Active Perception: Image Classification using Deep Reinforcement Learning" (2019, 7 citations), introduces a novel three-layer framework—comprising a meta-layer for goal-setting, an action layer for movement, and a perception layer for classification—that allows a robot to actively explore its environment to improve image recognition. This architecture addresses a critical challenge in robotics: how to plan and perceive when only incomplete information is available. By structuring decision-making hierarchically, Yuan’s approach enhances an agent's ability to dynamically adapt its observations, bridging the gap between reinforcement learning and real-world perception tasks. His contributions are particularly valuable for applications in autonomous systems, where efficient, goal-driven sensing is essential. With growing interest in active perception and embodied AI, Yuan’s work provides a foundational framework for future research in intelligent, resource-constrained agents.
Research Focus
Key Achievements
Top Papers
- 1