Hong-Xing Yu

Title: Innovations of Hong-Xing Yu in Dynamic Modeling

Introduction

Hong-Xing Yu is an accomplished inventor based in Stanford, CA. He has made significant contributions to the field of dynamic modeling and manipulation of multi-object scenes. His innovative approach combines advanced techniques in neural implicit scattering functions with model-predictive control frameworks.

Latest Patents

Hong-Xing Yu holds a patent titled "Neural implicit scattering functions for inverse parameter estimation and dynamics modeling of multi-object interactions." This patent describes a method for dynamic modeling and manipulation of multi-object scenes. The method utilizes object-centric neural implicit scattering functions (OSFs) as object representations within a model-predictive control (MPC) framework. It also includes modeling per-object light transport to enable compositional scene re-rendering under varying lighting conditions and object rearrangement. Furthermore, the method applies inverse parameter estimation and graph neural network (GNN) dynamics models to estimate initial object poses and light positions in multi-object scenes. This innovative approach allows for the manipulation of objects perceived in these scenes.

Career Highlights

Throughout his career, Hong-Xing Yu has worked with notable organizations, including the Toyota Research Institute and Leland Stanford Junior University. His experience in these institutions has contributed to his expertise in dynamic modeling and control systems.

Collaborations

Hong-Xing Yu has collaborated with esteemed colleagues such as Sergey Zakharov and Stephen Tian. Their joint efforts have further advanced the research and applications of dynamic modeling techniques.

Conclusion

Hong-Xing Yu's innovative work in dynamic modeling and manipulation of multi-object scenes showcases his expertise and contributions to the field. His patent and collaborations reflect a commitment to advancing technology in this area.