Xiaopeng Wu

Title: Innovations by Xiaopeng Wu in User Conversion Prediction

Introduction

Xiaopeng Wu is an accomplished inventor based in Kenmore, WA (US). He has made significant contributions to the field of online advertising through his innovative patent. His work focuses on enhancing user conversion predictions, which is crucial for effective advertisement targeting.

Latest Patents

Xiaopeng Wu holds a patent titled "User conversion prediction using a multi-task model." This patent describes systems and techniques for predicting user conversions in online advertising. The approach involves processing input data associated with user and advertisement features through neural networks to generate embedding representations or feature cross representations. A multi-task layer is utilized to calculate probabilities related to various user actions, such as clicks, page views, sign-ups, or purchases. The system effectively addresses challenges like data sparsity and negative transfer, thereby improving conversion metrics.

Career Highlights

Xiaopeng Wu is currently employed at Snap Inc., where he applies his expertise in user conversion prediction. His innovative work has the potential to significantly enhance advertisement targeting strategies, making them more efficient and effective.

Collaborations

Xiaopeng has collaborated with notable colleagues, including Weizhi Li and Joseph William Robinson. Their combined efforts contribute to advancing the field of online advertising and user behavior analysis.

Conclusion

Xiaopeng Wu's contributions to user conversion prediction represent a significant advancement in online advertising technology. His innovative patent and work at Snap Inc. highlight the importance of data-driven approaches in enhancing advertisement effectiveness.