Yifei Yang's avatar
Google Scholar GitHub

Embodied AI · Human Motion Generation · Humanoid Control

Yifei Yang

Ph.D. Candidate in Computer Science at Singapore Management University, advised by Prof. Shengfeng He.

My long-term goal is to build agents with general embodied intelligence that can understand open-ended worlds, learn and improve through experience, and eventually reach or exceed human-level capabilities. My work to date develops the action foundation for this vision through controllable human and humanoid motion.

Before beginning my Ph.D., I was a Research Engineer at Shanghai Artificial Intelligence Laboratory, where I developed MMGeneration as its primary contributor and maintainer and contributed to MMagic and MMEngine within OpenMMLab.

Email: yangyfaker@gmail.com Singapore

Recent Updates

Publications

Selected published work.

Lagrangian Motion Fields teaser

Lagrangian Motion Fields for Long-Term Motion Generation

Yifei Yang*, Zikai Huang*, Chenshu Xu, Shengfeng He

IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

A long-horizon motion representation that treats joints as Lagrangian particles and condenses redundant frames into interpretable supermotions.

Paper

DragNoise point-based image editing results compared with DragDiffusion and DragGAN

Drag Your Noise: Interactive Point-based Editing via Diffusion Semantic Propagation

Haofeng Liu, Chenshu Xu, Yifei Yang, Lihua Zeng, Shengfeng He

Conference on Computer Vision and Pattern Recognition, 2024

Interactive point-based image editing through diffusion semantic propagation, without per-step latent optimization.

Paper Code

Experience

Shanghai Artificial Intelligence Laboratory

Research Engineer · Shanghai, China

Developed MMGeneration as its primary contributor and maintainer, contributed 60k+ lines to OpenMMLab, and served as a core contributor to MMagic and MMEngine.

SenseTime Research

Research Intern · Shenzhen, China

Built the first version of MMGeneration from scratch, unifying more than 20 generative adversarial network algorithms behind one training and evaluation interface.

Education

Singapore Management University

Ph.D. in Computer Science · Singapore

Advised by Prof. Shengfeng He. Presidential Doctoral Fellowship, 2026/2027.

Shandong University

M.Sc. in Mathematics · Jinan, China

Advised by Prof. Lilu Zhao; analytic number theory; thesis: On the Representation of Numbers as Sums of Squares and Cubes.

East China Normal University

B.Eng. in Computer Science · Shanghai, China

Open Source

MMGeneration

Primary contributor and maintainer. A toolbox for training and evaluating generative adversarial networks and diffusion models.

MMagic

Core contributor. An image and video generation, editing, restoration, and enhancement toolbox.

MMEngine

Core contributor. The foundational training engine shared across OpenMMLab algorithm libraries.

Honors

  • Singapore Management University Presidential Doctoral Fellowship, 2026/2027
  • Fancy Idea Award, Alibaba–Tsinghua Security AI Challenger Program, 2021
  • 8th of 953 teams, ZhengTu Machine Vision AI Competition, 2020

Teaching & Service

  • Reviewer, International Conference on Machine Learning, 2026
  • Teaching Assistant, Generative Models for Natural Language Processing and Multimodal Reasoning, 2025