Company Filing History:
Years Active: 2024
Title: Innovations of John Duchi
Introduction
John Duchi is a notable inventor based in Menlo Park, California. He has made significant contributions to the field of machine learning, particularly in the area of private federated learning. With a total of two patents to his name, Duchi's work is at the forefront of technological advancements.
Latest Patents
Duchi's latest patents include "User behavior model development with private federated learning." This invention provides a non-transitory machine-readable medium that stores instructions for processors to receive a machine learning model from a server. It detects usage patterns for content items and trains the model using local data to generate a trained machine learning model. The process includes generating and transmitting a privatized update for the machine learning model back to the server.
Another significant patent is "Private federated learning with protection against reconstruction." This invention also involves a non-transitory machine-readable medium that enables processors to receive a machine learning model at a client device. It trains the model using local data, generates an update that includes a weight vector, and privatizes this update before transmitting it to the server.
Career Highlights
John Duchi is currently employed at Apple Inc., where he continues to innovate and develop cutting-edge technologies. His work has been instrumental in enhancing the capabilities of machine learning applications, particularly in ensuring user privacy.
Collaborations
Duchi has collaborated with notable coworkers such as Abhishek Bhowmick and Julien F Freudiger. Their combined expertise contributes to the advancement of their projects and the overall success of their innovations.
Conclusion
John Duchi's contributions to machine learning and private federated learning are noteworthy. His patents reflect a commitment to advancing technology while prioritizing user privacy. Duchi's work at Apple Inc. continues to influence the future of machine learning applications.