Siyi Chen

me

Email: siyich@umich.edu

Google Scholar / Github / CV

[Research] [Teaching] [Project]

About Me

Hi, I am Siyi Chen (陈思怡). You can also call me Stella.

Currently, I am an ECE PhD student in University of Michigan-Ann Arbor (2022 - Present), luckily advised by Professor Qing Qu. My research interests encompass generative models, representation learning, and their connections. I am keen on understanding the learning mechanisms and the acquired information structure of generative and perceptual models. I aim to utilize the theoretical understandings for unified and interpretable applications while combining inspirations from optimization techniques, principles of physics, and human cognition. Specifically, I am exploring these questions in diffusion models, multimodal learning, self-supervised learning, and dynamics modeling.

Previously, I was a dual degree undergraduate student (2018 - 2022): At University of Michigan-Ann Arbor, I majored in CSE & minored in Mathematics. At UMJI, SJTU, I majored in ECE. At UM, I was luckily advised by professor David Fouhey and mentored by Shengyi Qian on computer vision.

News

  1. [10/2024] One paper on self-superised video representation learning is accepted by NeurIPS 2024 SSL Workshop!
  2. [10/2024] Two papers on diffusion models are accepted by NeurIPS 2024 M3L Workshop!
  3. [09/2024] Our work Exploring Low-Dimensional Subspaces in Diffusion Models for Controllable Image Editing is accepted by NeurIPS 2024!
  4. [09/2024] Two papers (paper1, paper2) on dffisuion models are released!
  5. [09/2024] One paper on self-superised video representation learning is released!
  6. [05/2024] I have passed my PhD qualification exam!

Publications

Exploring Low-Dimensional Subspaces in Diffusion Models for Controllable Image Editing (NeurIPS 2024)

Siyi Chen*, Huijie Zhang*, Minzhe Guo, Yifu Lu, Peng Wang, Qing Qu

[Paper] [Website] [Code]

Unfolding Videos Dynamics via Taylor Expansion (NeurIPS 2024 Workshop)

Siyi Chen, Minkyu Choi, Zesen Zhao, Kuan Han, Qing Qu, Zhongming Liu

[Paper]

Diffusion Models Learn Low-Dimensional Distributions via Subspace Clustering (NeurIPS 2024 Workshop)

Peng Wang, Huijie Zhang, Zekai Zhang, Siyi Chen, Yi Ma, Qing Qu

[Paper] [Code]

Understanding 3D Object Articulation in Internet Videos (CVPR 2022)

Shengyi Qian, Linyi Jin, Chris Rockwell, Siyi Chen, David Fouhey

[Paper] [Website] [Code]