Yi Chang

Papers

Total Citations

H-Index

About

Yi Chang is a rising researcher in artificial intelligence, with a primary focus on reinforcement learning (RL) and multi-agent systems. His work tackles fundamental challenges in training robust, generalizable agents. In his 2023 paper, "Learning Generalizable Agents via Saliency-Guided Features Decorrelation," Chang addresses a critical bottleneck in visual-based RL: the tendency of agents to overfit to task-irrelevant environmental variations like background noise. By introducing a saliency-guided decorrelation mechanism, he helps agents focus on task-relevant features, significantly improving generalization to unseen environments. This work has already garnered early impact with 3 citations. In 2022, Chang contributed "Distributional Reward Estimation for Effective Multi-Agent Deep Reinforcement Learning," which confronts the high reward uncertainty inherent in multi-agent settings—a key obstacle in applications like robotics and autonomous driving. By modeling the full distribution of rewards rather than just their expectation, his approach enables more stable and effective policy learning. With 2 citations, this work underscores his growing influence. Chang’s research is paving the way for more reliable and adaptable AI systems, making him a promising voice in the next generation of reinforcement learning researchers.