Hsin-Yi Lin

Papers

Total Citations

H-Index

About

Hsin-Yi Lin is a researcher in reinforcement learning and intelligent agent systems, with a focus on accelerating learning processes in complex environments. Her most-cited work, "A CMAC-Q-Learning based Dyna agent" (2008, 7 citations), introduces a novel integration of cerebellar model articulation controllers (CMAC), Q-learning, and prioritized sweeping techniques. This hybrid approach addresses the critical challenge of slow learning speed in reinforcement learning by combining function approximation with efficient planning strategies. Lin's contribution lies in demonstrating how CMAC's generalization capabilities can be effectively paired with Dyna architecture's model-based planning to shorten training time while maintaining learning quality. Her work is particularly relevant for applications requiring rapid adaptation, such as robotics and autonomous systems, where traditional Q-learning may converge too slowly. Though her citation count is modest, Lin's research represents an important step in making reinforcement learning more practical for real-time decision-making tasks. Her approach of merging neural-inspired function approximators with temporal-difference learning continues to influence subsequent work in accelerating reinforcement learning algorithms.