BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260417T190104Z
LOCATION:West Building\, Rooms 220-222
DTSTART;TZID=America/Los_Angeles:20250813T154500
DTEND;TZID=America/Los_Angeles:20250813T173500
UID:siggraph_SIGGRAPH 2025_sess156@linklings.com
SUMMARY:Physics-Based Human Characters
DESCRIPTION:Diffuse-CLoC: Guided Diffusion for Physics-based Character Loo
 k-ahead Control\n\nMeet Diffuse-CLoC—a powerful unification of intuitive s
 teering in kinematic motion generation and physics-based character control
 . By guiding diffusion over joint state-action spaces, it enables agile, s
 teerable, and physically realistic motions across diverse downstream tasks
 —from obsta...\n\n\nXiaoyu Huang (University of California Berkeley, Robot
 ics and AI Institute); Takara Truong (Stanford University, Robotics and AI
  Institute); Yunbo Zhang, Fangzhou Yu, Jean Pierre Sleiman, and Jessica Ho
 dgins (Robotics and AI Institute); Koushil Sreenath (Robotics and AI Insti
 tute, University of California Berkeley); and Farbod Farshidian (Robotics 
 and AI Institute)\n---------------------\nPolicy-Space Diffusion for Physi
 cs-Based Character Animation\n\nWe present a new perspective on physics-ba
 sed character animation. Assuming policies for similar motions should have
  similar weights, we introduce regularization during RL training to preser
 ve weight similarity. By modeling the weights’ manifold with a diffusion m
 odel, we generate a continuum ...\n\n\nMichele Rocca, Sune Darkner, and Ke
 nny Erleben (University of Copenhagen); Sheldon Andrews (École de Technolo
 gie Supérieure (ÉTS)); and Michele Rocca\n---------------------\nMAGNET: M
 uscle Activation Generation Networks for Diverse Human Movement\n\nWe intr
 oduce MAGNET (Muscle Activation Generation Networks), a scalable framework
  for reconstructing full-body muscle activations across diverse human move
 ments, which also includes distilled models for solving downstream tasks o
 r generating real-time muscle activations—even on edge devices. T...\n\n\n
 Jungnam Park, Euikyun Jung, Jehee Lee, and Jungdam Won (Seoul National Uni
 versity)\n---------------------\nPARC: Physics-based Augmentation with Rei
 nforcement Learning for Character Controllers\n\nPARC is a framework that 
 enhances terrain traversal with machine learning and physics-based simulat
 ion. By iteratively training a kinematic motion generator and simulated mo
 tion tracker, PARC produces a character controller capable of traversing c
 omplex environments using highly agile motor skills, ...\n\n\nMichael Xu, 
 Yi Shi, and KangKang Yin (Simon Fraser University) and Xue Bin Peng (Simon
  Fraser University, NVIDIA)\n---------------------\nViSA: Physics-based Vi
 rtual Stunt Actors for Ballistic Stunts\n\nWe introduce ViSA (Virtual Stun
 t Actors), an interactive animation system using deep reinforcement learni
 ng to generate realistic ballistic stunt actions. It efficiently produces 
 dynamic scenes commonly seen in films and TV dramas, such as traffic accid
 ents and stairway falls. A novel action space d...\n\n\nMinseok Kim and Wo
 njeong Seo (Seoul National University), Sung-Hee Lee (Korea Advanced Insti
 tute of Science and Technology (KAIST)), and Jungdam Won (Seoul National U
 niversity)\n---------------------\nPLT: Part-Wise Latent Tokens as Adaptab
 le Motion Priors for Physically Simulated Characters\n\nWe introduce a phy
 sically-based character animation framework that exploits part-wise latent
  tokens. The novel structured decomposition enables dynamic exploration to
  stably adapt to diverse unseen scenarios. Additional refinement networks 
 improve overall motion quality. We show superior performance...\n\n\nJinse
 ok Bae, Younghwan Lee, Donggeun Lim, and Young Min Kim (Seoul National Uni
 versity)\n---------------------\nPhysics-Based Human Characters - Interact
 ive Discussion\n\nAfter the summary presentations, attendees will particip
 ate in an interactive discussion. Outside the room will be a series of pos
 ter boards for authors to gather around with the audience. Authors are inv
 ited to bring any material related to their paper that could instigate fur
 ther conversation such...\n\n---------------------\nPhysicsFC: Learning Us
 er-Controlled Skills for a Physics-Based Football Player Controller\n\nPhy
 sicsFC introduces a breakthrough in interactive football simulation—enabli
 ng real-time control of physically simulated players that perform complex 
 skills with smooth transitions. It combines skill-specific learning, physi
 cs-informed rewards, latent-guided training, and transition-aware sta...\n
 \n\nMinsu Kim, Eunho Jung, and Yoonsang Lee (Hanyang University)\n\nIntere
 st Area: Research & Education\n\nRecording: Livestreamed, Not Livestreamed
 , Recorded, Not Recorded\n\nRegistration Category: Full Conference, Virtua
 l Access, Wednesday\n\nSession Chair: Yuting Ye (Reality Labs Research, Me
 ta; Meta)
END:VEVENT
END:VCALENDAR