Tensor completion for PDEs with uncertain coefficients and Bayesian Update Alexander Litvinenko (joint work with E. Zander, B. Rosic, O. Pajonk, H. Matthies) Center for Uncertainty Quantification http://sri-uq.kaust.edu.sa/ Extreme Computing Research Center, KAUST Alexander Litvinenko (joint work with E. Zander, B. Rosic, O. Pajonk, H. Matthies) Tensor completion for PDEs with uncertain coefficients and B
32
Embed
Tensor Completion for PDEs with uncertain coefficients and Bayesian Update techniques
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Tensor completion for PDEs with uncertaincoefficients and Bayesian Update
Alexander Litvinenko(joint work with E. Zander, B. Rosic, O. Pajonk, H. Matthies)
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
http://sri-uq.kaust.edu.sa/
Extreme Computing Research Center, KAUST
Alexander Litvinenko (joint work with E. Zander, B. Rosic, O. Pajonk, H. Matthies)Tensor completion for PDEs with uncertain coefficients and Bayesian Update
Part I (Stochastic forward problem):1. Motivation2. Elliptic PDE with uncertain coefficients3. Discretization and low-rank tensor approximations
Part II (Bayesian update):1. Bayesian update surrogate2. Examples
Part III (Tensor completion):1. Problem setup2. Tensor completion for Bayesian Update
4*
Motivation to do Uncertainty Quantification (UQ)
Motivation: there is an urgent need to quantify and reduce theuncertainty in output quantities of computer simulations withincomplex (multiscale-multiphysics) applications.
Typical challenges: classical sampling methods are often veryinefficient, whereas straightforward functional representationsare subject to the well-known Curse of Dimensionality.Nowadays computational predictions are used in criticalengineering decisions and thanks to modern computers we areable to simulate very complex phenomena. But, how reliableare these predictions? Can they be trusted?
Example: Saudi Aramco currently has a simulator,GigaPOWERS, which runs with 9 billion cells. How sensitiveare the simulation results with respect to the unknown reservoirproperties?
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
3 / 30
4*
Part I: Stochastic forward problem
Part I: Stochastic Galerkin method to solveelliptic PDE with uncertain coefficients
4*
PDE with uncertain coefficient and RHS
Consider− div(κ(x , ω)∇u(x , ω)) = f (x , ω) in G × Ω, G ⊂ R2,u = 0 on ∂G, (1)
where κ(x , ω) - uncertain diffusion coefficient. Since κ positive,usually κ(x , ω) = eγ(x ,ω).For well-posedness see [Sarkis 09, Gittelson 10, H.J.Starkloff11, Ullmann 10].Further we will assume that covκ(x , y) is given.
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
4 / 30
4*
My previous work
After applying the stochastic Galerkin method, obtain:Ku = f, where all ingredients are represented in a tensor format
Compute maxu, var(u), level sets of u, sign(u)[1] Efficient Analysis of High Dimensional Data in Tensor Formats,
Espig, Hackbusch, A.L., Matthies and Zander, 2012.
Research which ingredients influence on the tensor rank of K[2] Efficient low-rank approximation of the stochastic Galerkin matrix in tensor formats,
Wahnert, Espig, Hackbusch, A.L., Matthies, 2013.
Approximate κ(x , ω), stochastic Galerkin operator K in TensorTrain (TT) format, solve for u, postprocessing[3] Polynomial Chaos Expansion of random coefficients and the solution of stochastic
partial differential equations in the Tensor Train format, Dolgov, Litvinenko, Khoromskij, Matthies, 2016.
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
5 / 30
4*
Canonical and Tucker tensor formats
Definition and Examples of tensors
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
6 / 30
4*
Canonical and Tucker tensor formats
[Pictures are taken from B. Khoromskij and A. Auer lecture course]
Storage: O(nd )→ O(dRn) and O(Rd + dRn).
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
7 / 30
4*
Definition of tensor of order d
Tensor of order d is a multidimensional array over a d-tupleindex set I = I1 × · · · × Id ,
A = [ai1...id : iµ ∈ Iµ] ∈ RI , Iµ = 1, ...,nµ, µ = 1, ..,d .
A is an element of the linear space
Vn =d⊗µ=1
Vµ, Vµ = RIµ
equipped with the Euclidean scalar product 〈·, ·〉 : Vn ×Vn → R,defined as
〈A,B〉 :=∑
(i1...id )∈I
ai1...id bi1...id , for A, B ∈ Vn.
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
8 / 30
4*
Discretization of elliptic PDE
Now let us discretize our diffusion equation withuncertain coefficients
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
I κ(x, ξ): N = 100 dofs, M = 5, number of KLE terms 35, beta distribution for κ, Gaussian covκ, cov.length 0.1, multi-variate Hermite polynomial of order pκ = 2;
I RHS f (x, ξ): Mf = 5, number of KLE terms 40, beta distribution for κ, exponential covf , cov. length 0.03,multi-variate Hermite polynomial of order pf = 2;
I b.c. g(x, ξ): Mg = 2, number of KLE terms 2, normal distribution for g, Gaussian covg , cov. length 10,multi-variate Hermite polynomial of order pg = 1;
I pφ = 3 and pu = 3
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
22 / 30
4*
Example: updating of the solution u
0 0.5 1-20
0
20
40
60
0 0.5 1-20
0
20
40
60
Figure: Original and updated solutions, mean value plus/minus 1,2,3standard deviations
[graphics are built in the stochastic Galerkin library sglib, written by E. Zander in TU Braunschweig]
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
23 / 30
4*
Example: Updating of the parameter
0 0.5 10
0.5
1
1.5
0 0.5 10
0.5
1
1.5
Figure: Original and updated parameter κ.
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
24 / 30
4*
Part III. Tensor completion
Now, we consider how toapply Tensor Completion Techniques
for Bayesian Update
In Bayesian Update surrogate, the assimilated PCE coeffs ofparameter qa will beNEW gPCE coeffs=OLD gPCE coeffs + gPCE of UpdateALL INGREDIENTS ARE TENSORS!
Problem of fitting a low rank tensor A ∈ RI , I := I1 × ...× Id ,Iµ = 1, ...,nµ, µ ∈ D := 1, ..,d, to given data points
Mi ∈ R | i ∈ P, P ⊂ I, #P ≥d∑µ=1
nµ, (11)
by minimizing the distance between the given values (Mi)i∈Pand approximations (Ai)i∈P :
A = argminA∈T
∑i∈P
(Mi − Ai)2 (12)
Remark: here we assume that our target tensor M allows for alow rank approximation ‖M − M‖ ≤ ε, ε ≥ 0 and M fulfillscertain rank bounds, T - Low rank format under consideration.
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
26 / 30
4*
Problem setup: Tensor completion
L. Grasedyck et all, 2016, hierarchical and tensor train formatsW. Austin, T, Kolda, D, Kressner, M. Steinlechner et al, CPformat
Goal: Reconstruct tensor with O(log N) number of samples.Methods:1. ALS inspired by LMaFit method for matrix completion,complexity O(r4d#P).2. Alternating directions fitting (ADF), complexity O(r2d#P).
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
27 / 30
4*
Numerical experiments for SPDEs: Tensor completion
[L. Grasedyck, M. Kluge, S. Kraemer, SIAM J. Sci. Comput., Vol 37/5, 2016]
Applied ALS and ADF methods to:− div(κ(x , ω)∇u(x , ω)) = 1 in D × Ω,u(x , ω) = 0 on ∂G × Ω,
(13)
D = [−1,1]. The goal is to determine u(ω) :=∫
D u(x , ω)dx .FE with 50 dofs, KLE with d terms, d-stochastic independentRVs,Yields to tensor Ai1...id := u(i1, ..., id ),n = 100, d = 5, slice density CSD = 6.Software (matlab) is available.
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
28 / 30
4*
Example: updating of the solution u
0 0.5 1-20
0
20
40
60
0 0.5 1-20
0
20
40
60
0 0.5 1-20
0
20
40
60
0 0.5 1-20
0
20
40
60
0 0.5 1-20
0
20
40
60
Figure: Original and updated solutions, mean value plus/minus 1,2,3standard deviations. Number of available measurements 0,1,2,3,5
[graphics are built in the stochastic Galerkin library sglib, written by E. Zander in TU Braunschweig]
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up
29 / 30
4*
Conclusion
I Introduced low-rank tensor methods to solve elliptic PDEswith uncertain coefficients,
I Explained how to compute the maximum and the mean inlow-rank tensor format,
I Derived Bayesian update surrogate ϕ (as a linear,quadratic, cubic etc approximation), i.e. computeconditional expectation of q, given measurement y .
I Apply Tensor Completion method to sparse measurementtensor in the likelihood.
Center for UncertaintyQuantification
Center for UncertaintyQuantification
Center for Uncertainty Quantification Logo Lock-up