Xiatian Zhu

Title: Xiatian Zhu: Innovator in Generative Models

Introduction

Xiatian Zhu is a prominent inventor based in Shanghai, China. He has made significant contributions to the field of generative models, particularly through his innovative patent. His work is recognized for its potential to enhance the efficiency of score-based generative models.

Latest Patents

Xiatian Zhu holds a patent titled "Method and system for accelerating score-based generative models with preconditioned diffusion sampling." This patent describes a method for accelerating score-based generative models (SGM) by setting a frequency mask and a space mask, along with a target sampling iteration number. The process involves sampling an initial sample, conducting iterations that include sampling a noise term, applying a preconditioned diffusion sampling operator, and generating a preconditioned noise term. The method also includes calculating a drift term and diffusing the sample of each iteration, ultimately outputting the result. He has 1 patent to his name.

Career Highlights

Xiatian Zhu is affiliated with Fudan University, where he continues to engage in research and development in the field of generative models. His academic background and ongoing projects contribute to the advancement of technology in this area.

Collaborations

Xiatian Zhu has collaborated with notable colleagues, including Fang Liu and Hengyuan Ma. Their combined expertise enhances the research output and innovation at Fudan University.

Conclusion

Xiatian Zhu is a key figure in the realm of generative models, with a focus on improving their efficiency through innovative methods. His contributions are paving the way for future advancements in this exciting field.