Lichao Sun

Title: Lichao Sun: Innovator in Neural Networks and Privacy-Preserving Technologies

Introduction

Lichao Sun is a prominent inventor based in Chicago, IL (US). He has made significant contributions to the fields of neural networks and privacy-preserving technologies. With a total of 4 patents, his work focuses on enhancing the robustness of neural network models and ensuring data privacy during deep learning processes.

Latest Patents

One of Lichao Sun's latest patents is titled "Robustness evaluation via natural typos." This innovative method evaluates the robustness of target neural network models by utilizing natural typos. The process involves generating natural typos based on specific rules and testing the neural network's output against ground truth labels. This approach provides valuable insights into the performance of neural networks in real-world scenarios.

Another notable patent is "Near-zero-cost differentially private deep learning with teacher ensembles." This patent outlines a framework for differentially private deep learning that protects training data privacy. By transferring knowledge from an ensemble of teacher models to a student model, this method ensures that privacy guarantees are maintained while still allowing for effective learning.

Career Highlights

Lichao Sun is currently employed at Salesforce.com, Inc., where he continues to push the boundaries of innovation in technology. His work at Salesforce focuses on developing advanced solutions that integrate machine learning and data privacy.

Collaborations

Throughout his career, Lichao has collaborated with notable colleagues, including Caiming Xiong and Jia Li. These collaborations have further enriched his research and development efforts, leading to impactful advancements in the field.

Conclusion

Lichao Sun is a distinguished inventor whose work in neural networks and privacy-preserving technologies has made a significant impact. His innovative patents reflect his commitment to advancing technology while addressing critical issues related to data privacy and model robustness.