Trung Dang

Title: Trung Dang - Innovator in Federated Learning

Introduction

Trung Dang is a notable inventor based in Los Angeles, CA. He has made significant contributions to the field of machine learning, particularly in federated learning. His innovative work has led to the development of a patent that addresses critical aspects of model updates and predictions.

Latest Patents

Trung Dang holds a patent titled "Ascertaining and/or mitigating extent of effective reconstruction, of predictions, from model updates transmitted in federated learning." This patent focuses on determining how effectively predictions generated by a machine learning model can be reconstructed from model updates. These updates are derived from the predictions and involve applying specific loss techniques, such as cross-entropy loss. The implementations disclosed in the patent generate measures indicating the degree of conformity between reconstructions and predictions. These measures play a crucial role in deciding whether to utilize particular loss techniques in the federated learning process.

Career Highlights

Trung Dang is currently employed at Google Inc., where he continues to push the boundaries of innovation in machine learning. His work is instrumental in enhancing the efficiency and effectiveness of federated learning models. With a focus on practical applications, Trung's contributions are shaping the future of artificial intelligence.

Collaborations

Trung collaborates with talented individuals such as Om Dipakbhai Thakkar and Swaroop Indra Ramaswamy. Together, they work on advancing technologies that leverage machine learning for various applications.

Conclusion

Trung Dang is a pioneering inventor whose work in federated learning is making a significant impact in the field of machine learning. His patent and collaborations reflect his commitment to innovation and excellence.