Physics-Informed Learning Machines for Multiscale and Multiphysics Problems

An image of various science-based visuals rolling out from a film reel labeled "philms."

(Image courtesy of Courtland Johnson | Pacific Northwest National Laboratory)

The PhILMs ("films") Center is a collaboration among Pacific Northwest National Laboratory and Sandia Laboratories, with academic partners at Brown University, Massachusetts Institute of Technology, Stanford University, and the University of California, Santa Barbara.

PhILMs is uncovering hidden physics based on new developments in deep learning, a computing technique that harnesses the power of machine learning and big data.

PhILMs investigators are developing physics-informed learning machines by encoding physics knowledge into deep learning networks to:

  • Design functional materials with tunable properties.
  • Solve longstanding problems in combustion, subsurface and earth systems, all exhibiting scaling cascades.
  • Establish probabilistic scientific computing as a new discipline at the interface of computational mathematics, multifidelity data, information fusion, and deep learning.

This work is supported by the Applied Mathematics Program within the U.S. Department of Energy Office of Advanced Scientific Computing Research.

PhILMs Latest News and Features

Upcoming Webinars

monday, June 14, 2021 at 1:00 PM PT/4:00 PM ET

"Galerkin Neural Networks: A Framework for Approximating Variational Equations with Error Control"
Mark Ainsworth, Brown University

Upcoming Conferences

Reduced-order models; Approximation theory; Machine learning; Surrogates, Emulators and Simulators

Marta D'Elia, Sandia National Laboratories, is organizing the conference "RAMSES: Reduced-order models; Approximation theory; Machine learning; Surrogates, Emulators and Simulators", held June 7-10, 2021 in SISSA, International School for Advanced Studies, Trieste, Italy.

MS: Data Management and Machine Learning

Qizhi He will be a co-organizer of a mini-symposium on "MS: Data Management and Machine Learning" at the U.S.Association for Computational Mechanics (USACM) Thematic Conference on Meshfree and Novel Finite Element Methods with Applications held September 25-27, 2022 in Berkeley, CA.

Awards and Recognition

Karniadakis interviewed for deeponet features in quanta magazine and techxplore.com

George Em Karniadakis talks about his DeepONet research in two new articles from Quanta Magazine and TechXplore.

Karniadakis Wins Award

George Em Karniadakis received the 2021 SIAM/ACM Prize in Computational Science and Engineering. The prize was awarded at the 2021 SIAM Conference on Computational Science and Engineering held virtually March 1-5, 2021.

Trask wins Early Career Award

Congratulations to Nat Trask, Sandia National Laboratories, who recently won an Early Career Research Program award from the U.S. Department of Energy for his proposed project on machine learning.

D'Elia Supports Online Project

Marta D'Elia has joined the committee of the One Nonlocal World project, an online platform where people working on nonlocal problems or interested in learning more about this topic can regularly check for updates on nonlocal modeling, analysis and computation. Nonlocal calculus and modeling is one of the focus areas in PhILMs.