DiJia Su

Title: DiJia Su: Innovator in Robotic Learning

Introduction

DiJia Su is a prominent inventor based in Sunnyvale, California. He has made significant contributions to the field of robotic learning, particularly in mitigating biases that can affect the performance of agents in various applications.

Latest Patents

DiJia Su holds a patent titled "Mitigating delusional bias in deep Q-learning for robotic and/or other agents." This patent addresses the issue of delusional bias that can occur in function approximation Q-learning. The disclosed techniques involve training and utilizing a value network to mitigate this bias, which can be crucial for generating effective actions for agents, including robotic and software agents. The patent also explores methods such as using a soft-consistency penalty and a search framework over multiple Q-functions to enhance performance.

Career Highlights

DiJia Su is currently employed at Google Inc., where he continues to innovate and develop advanced solutions in the realm of artificial intelligence and robotics. His work is instrumental in pushing the boundaries of what is possible in machine learning and agent-based systems.

Collaborations

He collaborates with talented colleagues, including Tyler Lu and Boon Teik Ooi, who contribute to his projects and research endeavors.

Conclusion

DiJia Su's work in mitigating delusional bias in deep Q-learning represents a significant advancement in robotic learning. His contributions at Google Inc. and his innovative patent highlight his role as a leading inventor in the field.