York, CA, United States of America

Borislav Mavrin

This inventor holds 1 USPTO granted patent. Top assignee: Huawei Technologies Co., Limited. Active years: 2022.


% Patents Active = 100.0

Average Co-Inventor Count = 2.0

ph-index = 1

Forward Citations = 3(Granted Patents)


Company Filing History:


Years Active: 2022

Loading Chart...
1 patent (USPTO):Explore Patents

Title: Borislav Mavrin: Innovator in Robotic Learning Systems

Introduction

Borislav Mavrin is a notable inventor based in York, CA (US). He has made significant contributions to the field of robotics, particularly in the area of reinforcement learning. His innovative approach focuses on enhancing the capabilities of robots to learn and adapt to various tasks.

Latest Patents

Mavrin holds a patent for "Systems and methods for learning reusable options to transfer knowledge between tasks." This invention involves a robot equipped with a reinforcement learning (RL) agent that learns to maximize cumulative rewards across tasks. The RL agent identifies minimally correlated features to create pseudo-rewards, known as feature rewards. Each feature reward corresponds to an option policy or skill that the RL agent learns to optimize. The agent is designed to select relevant features to develop respective option policies, ultimately enabling it to learn a new policy for different tasks by utilizing the learned options.

Career Highlights

Borislav Mavrin is currently employed at Huawei Technologies Co., Limited, where he continues to push the boundaries of robotic technology. His work emphasizes the importance of machine learning in developing intelligent systems that can efficiently transfer knowledge between tasks.

Collaborations

Mavrin collaborates with Daniel Mark Graves, contributing to advancements in their field through shared expertise and innovative ideas.

Conclusion

Borislav Mavrin's work in robotic learning systems exemplifies the potential of reinforcement learning in enhancing robotic capabilities. His contributions are paving the way for more intelligent and adaptable machines in the future.

Profile summary based on public USPTO records.
Please report any incorrect information to [email protected]
Loading…