Home

Welcome to my academic website! I am Xiaoyu Xie, a fifth-year PhD candidate in Mechanical Engineering Department at Northwestern University. I am supervised by Prof. Wing Kam Liu. My research focuses on developing robust and flexible artificial intelligence (AI) algorithms to discover hidden patterns and knowledge from data in manufacturing, fluid mechanics, solid mechanics, and more complex systems. My research goal is to improve the interpretability, generalization, and small-dataset scalability of AI algorithms by incorporating fundamental scientific insights. My research in dimensionless learning is a prime example of this approach, as it incorporates fundamental physical invariance to discover multiple levels of knowledge through machine learning.

News

• [06/2024]: I did a three-month internship at Amazon Annapurna Labs for optimizing distributed inference for multimodal deep learning models.
• [02/2024]: My internship paper about operator learning was released on Arxiv (Smooth and Sparse Latent Dynamics in Operator Learning with Jerk Regularization)
• [06/2023]: I did a three-month Research Internship at Mitsubishi Electric Research Laboratories (MERL).
• [03/2023]: I was invited to give a talk about Dimensionless Learning on the CMU SciML Webinar. (Recording)
• [11/2022]: I have been organizing a weekly AI and Science Reading Group affiliated with the Northwestern Institute on Complex Systems (NICO).
• [12/2022]: Our paper called “Data-driven discovery of dimensionless numbers and governing laws from scarce measurements“ was published on Nature Communications!
• [08/2022]: Our team won three 1st places and two 2nd places in AM-Bench Challenge, NIST, US!
• [06/2021]: Our work was published on npj Computational Materials-Nature!

eBook: Hands-on Scientific Machine Learning

I created a online interactive eBook for Scientific machine learning. You can check the details on this website. Each hands-on tutorial includes code, visualization, and youtube video for paper and implementation. You can also directly run all toturials in Colab or your local computers.

Research Interests:

• Scientific machine learning, AI for science;
• Data-driven reduced-order modeling;
• Data-driven scientific discovery;
• Machine learning-based digital twin;
• Computer vision;

Education

• Northwestern University, Evanston, IL, US
  • Ph.D. candidate in Mechanical Engineering, Sep. 2020 – Present
  • Advisor: Prof. Wing Kam Liu
• University of Chinese Academy of Sciences, Beijing, China
  • Inistitute of Mechanics
  • M.S. in General and Fundamental Mechanics, Sep. 2014 – Jul. 2017