Stephan Zheng

Title: Innovations of Inventor Stephan Zheng

Introduction

Stephan Zheng is an accomplished inventor based in Redwood City, CA. He has made significant contributions to the field of machine learning and reinforcement learning. With a total of 3 patents, his work focuses on solving complex tasks and enhancing learning systems.

Latest Patents

One of his latest patents is titled "Solving sparse reward tasks using self-balancing shaped rewards." This invention involves approaches for using self-balancing shaped rewards, which include randomly selecting a start and goal state. The method traverses first and second trajectories to move from the start state toward the goal state. The first terminal state of the first trajectory is closer to the goal state than the second terminal state of the second trajectory. The invention updates rewards for both trajectories using a self-balancing reward function based on the terminal states of the other trajectory. It also determines a gradient for the goal-oriented task module and updates parameters based on this gradient.

Another notable patent is "Learning world graphs to accelerate hierarchical reinforcement learning." This invention provides systems and methods for learning world graphs to enhance hierarchical reinforcement learning (HRL) for training machine learning systems. The approach includes a two-stage framework that consists of unsupervised world graph discovery and accelerated hierarchical reinforcement learning by integrating the graph.

Career Highlights

Stephan Zheng is currently employed at Salesforce Com, Inc., where he continues to innovate and develop new technologies. His work has had a significant impact on the field of machine learning, particularly in reinforcement learning applications.

Collaborations

He has collaborated with notable colleagues, including Alexander Richard Trott and Wenling Shang, contributing to various projects and advancements in their respective fields.

Conclusion

Stephan Zheng's innovative work and patents reflect his dedication to advancing technology in machine learning and reinforcement learning. His contributions are paving the way for future developments in these critical areas.