The goal of Lu Group’s research is to model and simulate physical and biological systems at different scales by integrating modeling, simulation, and machine learning, and to provide strategies for system learning, prediction, optimization, and decision making in real time. Our current research interest lies in scientific machine learning (SciML) and artificial intelligence for science (AI4Science), including theory, algorithms, software, and its applications to engineering, physical, and biological problems. Our broad research interests focus on multiscale modeling and high performance computing for physical and biological systems.

Ph.D., Master’s, & Undergraduate Students and Postdoc Opening: We are looking for Ph.D., Master’s, and undergraduate students, and Postdocs to work on scientific machine learning (SciML) and artificial intelligence for science (AI4Science). Students in statistics, mathematics, physics, computer science, engineering, or related majors with proficient coding skills are welcome to apply. Please feel free to contact me with CV (and/or transcripts, sample publications) attached if you are interested. For more information, please check the Ph.D. programs in Statistics and Data Science and Applied Mathematics, and postdoctoral positions in Data Science (Job posting), Applied Mathematics, and Neuroscience.

Recent News

  • New paper on arXiv: Speeding up and reducing memory usage for scientific machine learning via mixed precision. (Jan. 31, 2024)
  • Lu Lu gave a talk on “Accurate, efficient, and reliable learning of deep neural operators” at Stanford University, Department of Mechanical Engineering, Mechanics and Computation Seminar. (Dec. 7, 2023)
  • NVIDIA Deep Learning Institute has launched the Science and Engineering Teaching Kit developed by George Karniadakis, Raj Shukla, and Lu Lu in collaboration with NVIDIA. This course is available for free download for universities from the NVIDIA websites (here and here). For more information, please check out the news. (Nov. 13, 2023)
  • Highlight on AAS Nova: DeepONet offers an easier, faster way to model planet-forming disks. See the paper in Astrophysical Journal Letters. (Oct. 9, 2023)