BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030656Z
LOCATION:Meeting Room S221\, Level 2
DTSTART;TZID=Asia/Hong_Kong:20251215T170200
DTEND;TZID=Asia/Hong_Kong:20251215T171300
UID:siggraphasia_SIGGRAPH Asia 2025_sess115_papers_1613@linklings.com
SUMMARY:CFC: Simulating Character-Fluid Coupling using a Two-Level World M
 odel
DESCRIPTION:Zhiyang Dou and Chen Peng (The University of Hong Kong), Xinyu
  Lu (Centre for Transformative Garment Production), Xiaohan Ye (University
  of Hong Kong), Lixing Fang (University of Massachusetts Amherst), Yuan Li
 u (HKUST), Wenping Wang (Texas A&M University), Chuang Gan (University of 
 Massachusetts Amherst), Lingjie Liu (University of Pennsylvania), and Taku
  Komura (The University of Hong Kong)\n\nHumans possess the ability to mas
 ter a wide range of motor skills, using which they can quickly and flexibl
 y adapt to the surrounding environment. Despite recent progress in replica
 ting such versatile human motor skills, existing research often oversimpli
 fies or inadequately captures the complex interplay between human body mov
 ements and highly dynamic environments, such as interactions with fluids. 
 In this paper, we present a world model for Character-Fluid Coupling (CFC)
  for simulating human-fluid interactions via two-way coupling. We introduc
 e a two-level world model which consists of a Physics-Informed Neural Netw
 ork (PINN)-based model for fluid dynamics and a character world model capt
 uring body dynamics under various external forces. This two-level world mo
 del adeptly predicts the dynamics of fluid and its influence on rigid bodi
 es via force prediction, sidestepping the computational burden of fluid si
 mulation and providing policy gradients for efficient policy training. Onc
 e trained, our system can control characters to complete high-level tasks 
 while adaptively responding to environmental changes. We also present that
  the fluid initiates emergent behaviors of the characters, enhancing motio
 n diversity and interactivity. Extensive experiments underscore the effect
 iveness of CFC, demonstrating its ability to produce high-quality, realist
 ic human-fluid interaction animations.\n\nRegistration Category: Full Acce
 ss, Full Access Supporter\n\nSession Chair: Bo Ren (TMCC, College of Compu
 ter Science, Nankai University)\n\n
END:VEVENT
END:VCALENDAR