TriFusion: Towards Embodied Intelligence Across Humans, Avatars, and Humanoid Robotics aims to catalyze interdisciplinary dialogue and innovation at the intersection of computer graphics, artificial intelligence, robotics, and cognitive science. As digital avatars and humanoid robots increasingly act as proxies and collaborators in both virtual and physical environments, bridging the simulation-to-reality (sim-to-real) gap becomes a critical challenge for achieving believable, reliable, and adaptive embodied behaviors. This workshop investigates the shared principles and enabling technologies that support the seamless transfer of perception, cognition, and motor skills across human users, digital surrogates, and robotic agents.
Key topics include sensorimotor learning across simulated and real embodiments, physically based and data-driven motion synthesis, neural and biomechanical modeling for animation and control, real-time avatar retargeting, cross-domain telepresence, and the integration of digital twins as testbeds for embodied cognition. By uniting researchers and practitioners from graphics, animation, and robotics, TriFusion seeks to establish foundational frameworks for unified embodiment, advance collaborative simulation platforms, and chart a research agenda that supports scalable, adaptive, and human-aligned intelligence across both real and synthetic agents.
Date: Dec 8, 2025
Time: 9:00 am - 5:00 pm JST
Venue: SIGGRAPH Asia 2025
Location: TBD (Conference Venue)
Contact: trifusion2025@outlook.com
TriFusion explores how computer graphics can drive the next generation of embodied intelligence, bridging the sim-to-real gap between humans, avatars, and humanoid robots.
As virtual characters and physical agents increasingly share tasks and environments, this workshop spotlights the graphics-driven foundations of seamless motion transfer, real-time retargeting, and physically based animation. Topics include cross-domain sensorimotor learning, digital twins for simulation and evaluation, and unifying visual, biomechanical, and neural models for believable, adaptive motion across both virtual and real embodiments.
By connecting the graphics, animation, and robotics communities, TriFusion aims to establish shared tools, representations, and benchmarks for scalable, human-aligned embodied intelligence.
A comprehensive full-day exploration of embodied intelligence across humans, avatars, and humanoid robotics
Sunday, October 8, 2025 | SIGGRAPH Asia 2025 | Hong Kong, China
Full-Day Workshop: Towards Embodied Intelligence Across Humans, Avatars, and Humanoid Robotics
| 09:00 - 09:15 | Opening & Welcome Workshop Introduction and Objectives Organizers: Ping Luo (HKU) & Xiaokang Yang (SJTU) |
| 09:15 - 10:00 | Keynote 1: Bridging Sim-to-Real Gaps Prof. Jingyi Yu (ShanghaiTech University) "Computer Vision and VR/AR for Embodied Intelligence" |
| 10:00 - 10:45 | Keynote 2: Avatar-Human Interaction Prof. Ziwei Liu (Nanyang Technological University) "Human-Centric AI and Advanced Avatar Technologies" |
| 10:45 - 11:00 | Coffee Break & Networking |
| 11:00 - 11:45 | Keynote 3: Embodied AI & Robotics Prof. Jiajun Wu (Stanford University) "Intelligent Agents and Embodied Learning Systems" |
| 11:45 - 12:30 | Paper Session 1: Human-Avatar Interaction 4 Selected Research Presentations (10 min each + Q&A) Session Chair: Yao Mu (SJTU) |
| 12:30 - 13:30 | Lunch Break Networking & Poster Session |
| 13:30 - 14:15 | Paper Session 2: Humanoid Robotics 4 Selected Research Presentations (10 min each + Q&A) Session Chair: Yichao Yan (SJTU) |
| 14:15 - 15:00 | TriFusion Challenge Finals Track 1: Universal Motion Tracking Track 2: Visuo-Tactile Simulation Competition Results & Awards |
| 15:00 - 15:15 | Coffee Break & Poster Viewing |
| 15:15 - 16:00 | Interactive Demonstrations Live demos: Avatar-Robot Motion Transfer Visuo-Tactile Manipulation Systems Hands-on Experience Stations |
| 16:00 - 16:45 | Panel Discussion "Future of Embodied Intelligence: Challenges & Opportunities" Keynote speakers + Industry experts |
| 16:45 - 17:00 | Closing Ceremony & Awards Best Paper & Challenge Awards Future Directions & Call to Action |
Submit your latest research on:
Participate in our competitions:
Connect with experts from:
Expert in computer vision, computational photography, and virtual/augmented reality technologies.
Leading researcher in computer vision, human-centric AI, and advanced avatar technologies.
Pioneer in embodied AI and robotics, focusing on intelligent agents and embodied learning.
Yitian Liu (SJTU), Yuzhang Li (TJU), Yibin Liu (NEU), Yuhao Zhang (SJTU), Wanxi Dong (SUSTech)
Connect with researchers, practitioners, and enthusiasts working on embodied intelligence.
Contact: trifusion2025@outlook.com