Zeliang Kan

Title: Zeliang Kan: Innovator in AI Privacy Protection

Introduction

Zeliang Kan is a notable inventor based in London, GB. He has made significant contributions to the field of artificial intelligence, particularly in the area of privacy protection for machine learning models. His innovative approach addresses critical concerns regarding the security and integrity of AI systems.

Latest Patents

Zeliang Kan holds a patent for the "Detection of privacy attacks on machine learning models." This invention involves a method where a plurality of queries are input into an AI model composed of multiple layers, including an input layer, an output layer, and at least one intermediate layer. During inference, activations are intercepted from these intermediate layers, and a distance-based similarity analysis is conducted to determine if the queries are attempting to manipulate the AI model in an undesired manner. The findings are then communicated to a consuming application or process, enhancing the security of AI systems.

Career Highlights

Zeliang Kan is currently associated with Hiddenlayer, Inc., where he continues to develop innovative solutions in AI. His work focuses on ensuring that machine learning models can operate securely and effectively, safeguarding them against potential privacy attacks.

Collaborations

Zeliang collaborates with talented individuals such as Hengrui Jia and Ravikumar Balakrishnan. Their combined expertise contributes to advancing the field of AI and enhancing the robustness of machine learning technologies.

Conclusion

Zeliang Kan's contributions to AI privacy protection through his patent and work at Hiddenlayer, Inc. highlight his role as a key innovator in the field. His efforts are crucial in addressing the challenges posed by privacy attacks on machine learning models.