Physics Informed Machine Learning

A revolution in statistics and machine learning (ML) is underway. Modern algorithms can now learn high level abstractions via hierarchical models, leading to breakthrough accuracies in benchmarks for computer vision, language, etc. Underlying these advances is a strong and deep connection to various aspects of statistical physics. For example, classical coarse graining concepts such as the renormalization group directly map to deep learning. Another connections physics inspired algorithms for accelerated graphical model inference; originally designed for simulations of lattice models of magnetism and quantum field theory, these algorithms have proven transformative in the training of complex, hierarchical ML models.

This workshop seeks perspectives on leveraging the deep connection between ML and physics, but now with the goal to better understand and model physical systems, static and dynamic. We invite experts both in machine learning techniques as well as domain science applications such as building reduced models for infrastructures (energy systems, traffic flows, etc), emulating turbulent ids that arise in climate simulation, and reconstructing (from measurements and models) transport phenomena in complex materials. The workshop discussions are aimed towards approaches and methods for physical modeling applications where a big-data, black-box approach to ML is only a starting point. We seek participants who may suggest innovative approaches that extend application agnostic ML techniques by incorporating complex constraints imposed by physical principles (e.g. conservation laws, causality, entropy principles and related).

The workshop format will include lecture sessions, discussion sessions on applications, and posters. 

We plan for active participation of LANL researchers and program managers across directorates and divisions interested in the physics informed learning. This emerging area of research has many aspects of computational co-design,and draws on LANL's strengths in statistical physics, theoretical and applied computer science, infrastructure modeling and simulations, fluids and materials modeling, and high performance computing. Looking forward, we view physics informed learning as a viable path for LANL and DOE toward truly predictive multi-scale modeling, which is a foundational challenge for mechanical, materials,biological, and chemical engineering.

Topics:

• Energy
• Climate
  
• Materials
  
• Images