Yifan Wu

Hi! I am currently an intern at Tencent, Hunyuan on video world model research. Previously, I interned at Timi Studio Group, working on video generation research. I served as a research assistant at the CGVU Lab at The University of Hong Kong, supervised by Prof. Taku Komura. Before that, I received my M.S. degree from Boston University and my B.Eng. degree with honors from Yangzhou University. My research investigates the scalable, first-principles foundations of the world model paradigm—developing video-free architectures that learn the intrinsic dynamics of physical environments and leveraging inverse dynamics to enable general-purpose robotic intelligence. I am seeking Ph.D. opportunities starting Fall 2027.

Google Scholar GitHub Twitter LinkedIn

Education

  • Boston University

    M.S. in Electrical and Computer Engineering

  • Yangzhou University

    B.ENG. in Electrical Engineering

Honors & Awards

  • Tencent Project Up Scholarship Candidate
  • MIT-Interdisciplinary Quantum Information Science and Engineering
  • President's Excellence Scholarship
  • National College Students Mathematical Contest in Modeling, second place
  • 2023 ECE Master's Community Award

News

Experience

Tencent (Hunyuan)

Tencent (Hunyuan)

Intern

Tencent (Timi Studio Group)

Tencent (Timi Studio Group)

Intern

The University of Hong Kong

The University of Hong Kong

Research Assistant

SoftBank

SoftBank

Intern

ScaleAI

ScaleAI

Intern

Chinese Academy of Sciences

Chinese Academy of Sciences

Research Assistant

Selected Publications

View all

TrackFish3D: Self-Supervised 3D Tracking of Schooling Fish from Multi-view Videos

Patt Phurtivilai, Zhiyang Dou, Yifan Wu, Kinfung Chu, Yuan Liu, Lei Yang, Wenping Wang, Taku Komura

NeurIPS 2026

ID-Anchor:Identity-Preserving Character Animation via Tuning-Free Attention Masking for Multi-Person Scenes

Yifan Wu, Zhongfei Qin, Yuke Lou, Dafei Qin, Wenjia Wang, Zhiyang Dou, Peixu Xin, Songnan Li, Bo Hu, Taku Komura

Paper Submitted

4World:Hybrid World Model with Explicit 3D Memory and Generative Dynamics

Yuke Lou, Wenjia Wang, Bocheng Li, Yifan Wu, Zongyuan Yang, Baolin Liu, Taku Komura

Paper Submitted

EmbodMocap:In-the-Wild 4D Human-Scene Reconstruction for Embodied Agents

Wenjia Wang, Liang Pan, Huaijin Pi, Yuke Lou, Xuqian Ren, Yifan Wu, Zhongyingcheng Liao, Lei Yang, Taku Komura

CVPR 2026

SIMS:Simulating Human-Scene Interactions with Real World Script Planning

Wenjia Wang, Liang Pan, Zhiyang Dou, Jidong Mei, Zhouyingcheng Liao, Yuke Lou, Yifan Wu, Jingbo Wang, Lei Yang, Taku Komura

ICCV 2025

CBIL:Collective Behavior Imitation Learning for Fish from Real Videos

Yifan Wu, Zhiyang Dou, Yuko Ishiwaka, Shun Ogawa, Yuke Lou, Wenping Wang, Lingjie Liu, Taku Komura

SIGGRAPH Asia 2024 TOG (ACM Transactions on Graphics)

FBINeRF:Feature-Based Integrated Recurrent Network for Pinhole and Fisheye Neural Radiance Fields

Yifan Wu, Tianyi Cheng, Peixu Xin, Janusz Konrad

Course Project

Talks

Collective Behavior Imitation Learning and World Models

Sun Yat-sen University, Shenzhen Campus

Service

Conference Reviewer

ICCV, CVPR, SIGGRAPH

Journal Reviewer

ACM Transactions on Graphics (TOG)