Chiyu “Max” Jiang

'Max' Chiyu Jiang

Third year PhD student at UC Berkeley. Passionate programmer and applied mathematician. Interested in Machine Learning, AI, and its applications to fluid related physics.


Bio:
My name is Chiyu, and I go by ‘Max’​. I am currently a PhD student in UC Berkeley. I am interested in Deep Learning, Machine Learning, and its applications in fluids related controls problems. I also did several projects on shape generation and shape morphing. My research advisor is Professor Philip Marcus and I am co-adviced by Professor Phillip Colella in EECS Department / LBNL.

I came from a diverse personal as well as academic background. I was born in China, but I grew up living in several countries, namely China, Germany and Singapore, and I completed my undergraduate studies in Cornell University, Ithaca, NY in the United States. My hobbies include playing musical instruments (piano + guitar rules), photography, travel, and thinking about interesting scientific/entrepreneurial ideas.

Here’s a link to my personal blog and webpage. Currently showcasing a few of my research projects, but I’m hoping to put up some fun stuff as well:
http://www.maxjiang.ml

Current Research:

  • Deep Learning for fluid-rigid-body dynamics inference and model-based reinforcement learning for controls in fluid environment.
  • Airfoil design optimization with deep learning based flow inference.

Current Teaching (GSI):

Publications & Abstracts:

  1. Chiyu Jiang, Philip Marcus, et al. Hierarchical detail enhancing mesh-based shape generation with 3d generative adversarial network. arXiv preprint arXiv:1709.07581, 2017. Link
  2. Chiyu Jiang, Anzhu Sun, and Philip Marcus. Drag reduction of an airfoil using deep learning. Bulletin of the American Physical Society, 2017. Link
  3. Sahuck Oh, Chung-Hsiang Jiang, Chiyu Jiang, and Philip S. Marcus. Finding the optimal shape of the leading-and-trailing car of a high-speed train using design-by-morphing. Computational Mechanics, Oct 2017.
  4. Sahuck Oh, Chung-Hsiang Jiang, Philip Marcus, David Gutzwiller, Alain Demeulenaere, and Chiyu Jiang. Shape optimization of a turbine-99 draft tube using design-by-morphing. APS Division of Fluid Dynamics Meeting Abstracts, 2016. Link

Gallery:

3D Shape Generation using Deep Learning:



Left : Chairs and tables generated by proposed deep learning model, rendered by Blender;
Center: Morphing of two shapes by linear interpolation of latent space vector;
Right : Car generated by proposed deep learning model.

3D Mesh Morphing using Mesh Parameterization:



Left : Spherical parameterization of genus-zero mesh (horse) using conformal mean curvature flow (CMCF);
Center: Morphing of two mesh patches by boundary aligned mesh parametrization;
Right : Morphing of two genus-zero meshes by feature-aligned spherical parameterization.