Li Wan

Title: Innovations of Li Wan in Multimodal AI Security

Introduction

Li Wan is a prominent inventor based in Hangzhou, China. She has made significant contributions to the field of artificial intelligence, particularly in the area of security for multimodal AI models. Her innovative approach addresses critical challenges in ensuring the safety and integrity of AI systems.

Latest Patents

Li Wan holds a patent for a "Systematic security protection method for multimodal AI model, apparatus and device." This patent outlines a systematic security protection method that includes a security detection stage for the target multimodal AI model. The method utilizes a trained distance-aware judgment model to determine the perceptual distance between detection data and original data. It ensures that detection data meets specific requirements for effective security evaluation. If a security issue is identified, the method implements security coding protection and tokenization filtering, along with a multi-dimensional evaluation of the multimodal AI model.

Career Highlights

Li Wan is currently employed at Hangzhou Hikvision Digital Technology Co., Ltd. Her work at this leading technology company has allowed her to focus on developing advanced security solutions for AI applications. Her expertise in this domain has positioned her as a key player in the ongoing evolution of AI security measures.

Collaborations

Li Wan collaborates with talented colleagues, including Bin Wang and Kezhang Lin. These partnerships enhance her research and development efforts, fostering innovation in the field of AI security.

Conclusion

Li Wan's contributions to the field of multimodal AI security are noteworthy and impactful. Her innovative patent reflects her commitment to advancing technology in a safe and secure manner. As the landscape of artificial intelligence continues to evolve, her work will undoubtedly play a crucial role in shaping its future.