Abstract

Given a robust control system, simulation offers the potential for interactive humanoid characters that respond naturally to the actions of the user or changes in the environment. In recent years, we have seen significant progress in this area, where the physics-based humanoids become capable of reproducing a growing set of motion skills. In this talk, I will review recent progress in achieving these realistic and responsive human movements, with a focus on the approaches that utilize motion capture as reference. We will discuss the roles of abstract models, optimization, reinforcement learning, and deep (and not-so-deep) machine learning in those methods, as well the challenges that still need more efforts.

Biography

Libin Liu joined Peking University in July 2020 and is currently an assistant professor at Center on Frontiers of Computing Studies, PKU. Before joining CFCS, he was the Chief Scientist of DeepMotion Inc. He was a R&D postdoctoral associate at Disney Research and a postdoctoral research fellow at the University of British Columbia. He received his PhD in computer science from Tsinghua University in 2014. His research interests include character animation, physics-based simulation, motion control, and related areas such as optimization, reinforcement learning, and deep learning. He had served on the program committees of major conferences on computer graphics, including ACM SIGGRAPH, Eurographics, Pacific Graphics, etc.

Admission: https://www.wjx.cn/m/84200294.aspx
Online Talk: https://zoom.com.cn/j/95138086336