Skip to content
@Centrum-IntelliPhysics

Centrum IntelliPhysics

Our group aims to develop algorithms to accelerate traditional numerical solvers through the application of deep learning methodologies.

Welcome to Centrum IntelliPhysics Group at Johns Hopkins University

Like all good models, this page is iteratively updated. :)

We are an interdisciplinary research collective pioneering the convergence of physics-based modeling and scientific machine learning to tackle complex scientific and engineering challenges. “We don’t replace physics — we augment it with intelligent data-driven models.”


Research Themes

Our lab is driven by the vision of redefining scientific discovery by blending first-principles physics with data-driven intelligence. Our core research thrusts include:

  • Hybrid Physics-AI Solvers:
    Coupling neural operators with finite element methods to accelerate the time-to-solution as well as long rollouts for complex dynamics and multiscale modelin
  • Latent Representations for Physical Systems:
    Developing latent neural operators to learn low-dimensional, physics-consistent embeddings for design and control.
  • Scientific Machine Learning (SciML):
    Leveraging PINNs and physics-informed operator learning to accelerate discovery.
  • Mechanics Datasets:
    Compiling a collection of open-access mechanics datasets to evaluate the performance of machine learning models. Each dataset is generated using the open-source finite element software FEniCS and released under the CC BY-SA 4.0 license. You are welcome to download and use them for any research or educational purpose!

You can find a comprehensive list of our publications on our website. We also upload all our presentations from conferences and seminars on our GitHub Repository.

If you are interested in our work, you may also be interested in the SciML webinar series JHU-IITD SMaRT that we are hosting in collaboration with Prof. Souvik Chakraborty and Prof. Rajdip Nayek from the Indian Institute of Technology Delhi.

Popular repositories Loading

  1. Time-Marching-Neural-Operator-FE-Coupling Time-Marching-Neural-Operator-FE-Coupling Public

    This repository contains the code for coupling FEniCS with Physics-Informed DeepONet

    Python 19 3

  2. Conference-and-Seminar-Presentations Conference-and-Seminar-Presentations Public

    This repository contains the slides from the conference or seminar presentation of the group members

    6 1

  3. GNS_vs_NOs GNS_vs_NOs Public

    Comparing generalization performance of GNS against NOs for solving time-dependent PDEs.

    Python 5 1

  4. DeepONet-Efficient-Training-with-Random-Sampling DeepONet-Efficient-Training-with-Random-Sampling Public

    Efficient Training of Deep Neural Operator Networks via Randomized Sampling

    Jupyter Notebook 4 3

  5. PDEControl_DPC PDEControl_DPC Public

    Jupyter Notebook 4

  6. Neural-Operators-for-Natural-Hazards Neural-Operators-for-Natural-Hazards Public

    Neural operators for modeling stochastic response to seismic and wind excitations.

    Jupyter Notebook 3 2

Repositories

Showing 10 of 19 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…