A surrogate model is built to accelerate computationally expensive physical simulations, which is useful in multiquery problems, such as inverse problem, uncertainty quantification, design optimization, and optimal control. In this talk, two types of datadriven surrogate modeling techniques will be discussed, i.e., the blackbox approach that incorporates only data and the physicsinformed approach that incorporates the physics information as well as data within the surrogate models. The advantages and disadvantages of each method will be discussed. Furthermore, several recent developments of datadriven physicsinformed surrogate modeling techniques at LLNL will be introduced in the context of various physical simulations. For example, the reduced order model overcomes the difficulty of shock propagation phenomenon, achieving a speedup of O(2~10) with a relative error less than 1% for relatively small Lagrangian hydrodynamics problems. The space–time reduced order model accelerates largescale Neutron transport simulations by a factor of 7,000 with a relative error less than 1%. Finally, the nonlinear manifold reduced order model shows perfect marriage between machine learning and physicsinformed surrogate modeling and also solves the challenge imposed by the advectiondominated physical simulations.
You are here
Where are we with datadriven surrogate modeling for various physical simulations?
Research Group:
Youngsoo Choi
Institution:
Lawrence Livermore National Laboratory, CA
Schedule:
Friday, March 19, 2021  18:00
Location:
Online
Location:
Zoom Meeting
Abstract:
Openings
 Public Calls for Professors
 Temporary Professors/Researchers/Visiting Professors
 SISSA Mathematical Fellowships
 PostLauream Fellowships
 Research Training Fellowships
 Marie SklodowskaCurie Grants
 Open positions in MathLab
 Post Doctoral Fellowships
 PhD Scolarships
 SIS Fellowships
 Undergraduate Fellowships
 Postgraduate Fellowships
 MSc in Mathematics
 MSc in Data Science and Scientific Computing (DSSC)
 Professional Master Courses
 SISSA Mathematics Medals
Upcoming events

Solomiya Mizyuk
On the minimal models of surfaces via Mori theory
Friday, May 27, 2022  16:30 to 17:30
Today's Lectures

09:00 to 11:00

11:00 to 13:00

11:00 to 13:00

14:15 to 15:00

16:00 to 18:00
Recent publications

G.P. Leonardi; G. Saracco,Rigidity and trace properties...

M.W. Hess; A. Quaini; G. Rozza,DataDriven Enhanced Model Red...

M.W. Hess; A. Quaini; G. Rozza,A DataDriven Surrogate Modeli...

R. Della Marca; Mda Piedade Ramos; C. Ribeiro; A.Jacinta Soares,Mathematical modelling of osci...