Fanzi Wu

Title: Fanzi Wu: Inventor in Privacy-Preserving Machine Learning

Introduction

Fanzi Wu is an inventor based in Bothell, Washington. His work includes methods for training machine-learning models while protecting the privacy of individuals whose data is used.

Latest Patents

Wu is associated with 1 patent, titled “Differentially Private Open-Classification Machine Learning Model Training with Adaptive Clipping.” The patent describes methods and systems that limit the noise and influence contributed by individual members of a training dataset. It also describes adaptively adjusting those limits to reflect privacy needs while maintaining model performance, including in open-classification training.

Career Highlights

Wu works at Amazon Technologies, Inc. His patent focuses on balancing privacy protections with the performance of machine-learning models.

Collaborations

Wu’s listed coworker is Jun Fang.

Conclusion

Fanzi Wu’s patent work addresses adaptive privacy protections for machine-learning training, connecting data privacy with the practical performance of open-classification models.