Thursday, October 15, 2020

UQSay #16


The sixteenth UQSay seminar on Uncertainty Quantification and related topics, organized by L2S, MSSMAT, LMT and EDF R&D, will take place online on Thursday afternoon, October 22, 2020.

14h–15h — Nicolas Bousquet (EDF R&D)


Well-posed stochastic inversion in uncertainty quantification, with links with sensitivity analysis

Stochastic inversion problems are typically encountered when it is wanted to quantify the uncertainty affecting the inputs of computer models. They consist in estimating input distributions from noisy, observable outputs, and such problems are increasingly examined in Bayesian contexts where the targeted inputs are affected by a mixture of aleatory and epistemic uncertainties. While they are characterized by identifiability conditions, well-posedness constraints of "signal to noise" have to be took into account within the definition of the model, prior to inference. In addition to numeric conditioning notions and regularization techniques used in inverse problems, we propose and investigate an interpretation of well-posedness, in the context of parametric uncertainty quantification and global sensitivity analysis, based on the degradation of Fisher information. It offers an explicitation of such prior constraints considering linear or linearizable operators, this linearization being either local (based on differentiability) or variational. Simulated experiments indicate that, when injected into the modeling process, these constraints can limit the influence of measurement or process noise on the estimation of the input distribution, and let hope for future extensions in a full non-linear framework, for example through the use of linear Gaussian mixtures.​

Organizing committee: Julien Bect (L2S), Emmanuel Vazquez (L2S), Didier Clouteau (MSSMAT), Filippo Gatti (MSSMAT), Fernando Lopez Caballero (MSSMAT), Amélie Fau (LMT), Bertrand Iooss (EDF R&D).

Practical details: the seminar will be held online using Microsoft Teams.

If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay group on Teams, simply send an email and you will be invited. Please specify which email address the invitation must be sent to (this has to be the address associated with your Teams account).

You will find the link to the seminar on the "General" UQSay channel on Teams, approximately 15 minutes before the beginning.

The technical side of things: you can use Teams either directly from you web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.

Friday, October 2, 2020

UQSay #15


The fifteenth UQSay seminar on Uncertainty Quantification and related topics, organized by L2S, MSSMAT, and EDF R&D, will take place online on Thursday afternoon, October 8, 2020.

14h–15h — Sebastian Schöps (TU Darmstadt)


Uncertainty Quantification for Maxwell's eigenproblem based on isogeometric analysis and mode tracking

Superconducting cavities are used in particle accelerators, e.g. at DESY in Hamburg, Germany. Their resonating electromagnetic field is commonly characterised by eigenmodes and eigenvalues which are very sensitive to small geometry deformations. This presentation proposes an uncertainty quantification workflow based on a Karhunen–Loève expansion of the manufacturing imperfections and eigenvalue tracking based on algebraic and geometric homotopies.

Joint work with Niklas Georg, Wolfgang Ackermanna, Jacopo Corno.

Reference: DOI:10.1016/j.cma.2019.03.002 (arxiv:1802.02978).

Organizing committee: Julien Bect (L2S), Emmanuel Vazquez (L2S), Didier Clouteau (MSSMAT), Filippo Gatti (MSSMAT), Fernando Lopez Caballero (MSSMAT), Bertrand Iooss (EDF R&D).

Practical details: the seminar will be held online using Microsoft Teams.

If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay group on Teams, simply send an email and you will be invited. Please specify which email address the invitation must be sent to (this has to be the address associated with your Teams account).

You will find the link to the seminar on the "General" UQSay channel on Teams, approximately 15 minutes before the beginning.

The technical side of things: you can use Teams either directly from you web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.

Monday, September 21, 2020

UQSay #14


The fourteenth UQSay seminar on Uncertainty Quantification and related topics, organized by L2S, MSSMAT, and EDF R&D, will take place online on Thursday afternoon, September 24, 2020.

14h–15h — Amélie Fau (LMT, ENS Paris-Saclay)

Alternative strategies for adaptive sampling for kriging metamodels

A large variety of strategies have been proposed in the literature to offer optimal dataset for kriging metamodels. Even though adaptive schemes guarantee convergence and improvement of estimation accuracy for instance for Galerkin approaches at least in a goal-oriented sense, using usual adaptive sampling schemes for kriging metamodels might be detrimental, worsing prediction results compared to one-shot sampling techniques. The goal of this seminar is to share our experience on cases leading to this disvantageous behavior. Besides, problems leading to beneficial behavior will be discussed to highlight criteria for deciding about cases of interest for which adaptive sampling strategies are highly promising.

Joint work with Jan Fuhg & Udo Nackenhorst (Leibniz Universität, Hannover).

References: DOI:10.1007/s11831-020-09474-6 & the associated github repository.

Organizers: Julien Bect (L2S), Emmanuel Vazquez (L2S), Didier Clouteau (MSSMAT), Filippo Gatti (MSSMAT), Fernando Lopez Caballero (MSSMAT), Bertrand Iooss (EDF R&D).

Practical details: the seminar will be held online using Microsoft Teams.

If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay group on Teams, simply send an email and you will be invited. Please specify which email address the invitation must be sent to (this has to be the address associated with your Teams account).

You will find the link to the seminar on the "General" UQSay channel on Teams, approximately 15 minutes before the beginning.

The technical side of things: you can use Teams either directly from you web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.

Thursday, September 3, 2020

UQSay #13


The thirteenth UQSay seminar on Uncertainty Quantification and related topics, organized by L2S, MSSMAT, and EDF R&D, will take place online on Thursday afternoon, September 10, 2020.

14h–15h — Balázs Kégl (Noah's Ark Lab, Huawei Paris) — [slides]

DARMDN: Deep autoregressive mixture density nets for dynamical system modelling

Unlike computers, physical engineering systems (such as data center cooling or wireless network control) do not get faster with time. This is arguably one of the main reasons why recent beautiful advances in deep reinforcement learning (RL) stay mostly in the realm of simulated worlds and do not immediately translate to practical success in the real world. In order to make the best use of the small data sets these systems generate, we develop data-driven neural simulators to model the system and apply model-based control to optimize them. In this talk I will present the first step of this research agenda, a new versatile system modelling tool called deep autoregressive mixture density net (DARMDN – pronounced darm-dee-en). We argue that the performance of model-based reinforcement learning is partly limited by the approximation capacity of the currently used conditional density models and show how DARMDN alleviates these limitations. The model, combined with a random shooting controller, establishes a new state of the art on the popular Acrobot benchmark. Our most interesting and counter-intuitive finding is that the “sincos” Acrobot system which requires no multimodal posterior predictives, can be solved with a deterministic model, but only if it is trained as a probabilistic model. A deterministic model that is trained to minimize MSE leads to prediction error accumulation.

Joint work with Gabriel Hurtado and Albert Thomas.

Organizers: Julien Bect (L2S), Emmanuel Vazquez (L2S), Didier Clouteau (MSSMAT), Filippo Gatti (MSSMAT), Fernando Lopez Caballero (MSSMAT), Bertrand Iooss (EDF R&D).

Practical details: the seminar will be held online using Microsoft Teams.

If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay group on Teams, simply send an email and you will be invited. Please specify which email address the invitation must be sent to (this has to be the address associated with your Teams account).

You will find the link to the seminar on the "General" UQSay channel on Teams, approximately 15 minutes before the beginning.

The technical side of things: you can use Teams either directly from you web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.

Friday, August 7, 2020

UQSay will be back in September

Dear colleagues, friends, UQ lovers,

UQSay seminars will be back soon, on Thursday afternoons (2–3 PM) as usual:

UQSay #13 — September 10 — Balazs Kegl (Huawei Research)
DARMDN: Deep autoregressive mixture density nets for dynamical system modelling

UQSay #14 — September 24 — Amélie Fau (ENS Paris-Saclay / LMT)
(To be announced)

UQSay #15 — October 8 — Sebastian Schöps (TU Darmstadt)
Uncertainty Quantification for Maxwell's eigenproblem based on isogeometric analysis and mode tracking

UQSay #16 — October 22 — Nicolas Bousquet (EDF R&D)
Well-posed stochastic inversion in uncertainty quantification, with links with sensitivity analysis

☞☞☞  Save the dates!  ☜☜☜

Organizers: Julien Bect (L2S), Emmanuel Vazquez (L2S), Didier Clouteau (MSSMAT), Filippo Gatti (MSSMAT), Fernando Lopez Caballero (MSSMAT), Bertrand Iooss (EDF R&D).

Practical details: at least the first three seminars will be held online using Microsoft Teams.

If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay group on Teams, simply send an email and you will be invited. Please specify which email address the invitation must be sent to (this has to be the address associated with your Teams account).

You will find the link to the seminar on the "General" UQSay channel on Teams, approximately 15 minutes before the beginning.

The technical side of things: you can use Teams either directly from you web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.

Sunday, June 28, 2020

UQSay #12


The twelfth UQSay seminar on Uncertainty Quantification and related topics, organized by L2S, MSSMAT, and EDF R&D, will take place online on Thursday afternoon, July 9, 2020.

14h–15h — Roberto Miorelli (CEA Paris-Saclay LIST) — [slides]

Metamodels and Statistical Tools Applied to Non-destructive Testing Problems

In the last decades, the non-destructive testing (NDT) research community has greatly increased its interest on statistical methods. This interest was driven by the community willingness to provide efficient solutions to NDT problems and push further the use of simulation to better design NDT inspection systems and procedures.

This research brought to the developments of new solutions to carry out sensitivity analysis, compute model assisted probability of detection, perform flaw(s) detection, flaw(s) characterization and probe optimization. Toward this end, the use of metamodels (also known as surrogate models) have been widely exploited for different NDT physics (ultrasound testing, eddy current testing, infrared thermography, etc.) with promising results. More recently, the NDT community is focussing on providing advanced simulation tools for performance estimation under uncertainties. One of the main area of interest of this research turns concerns the estimation of uncertainties for flaw(s) detection, characterization, and optimization problems. These developments have been particularly boosted and supported by the recent achievements in the field of machine learning that provide a plethora of algorithms and tools suitable for solving many different NDT problems.

This talk provides an overview on the use of metamodel tools and statistical methods applied to NDT problems developed at CEA LIST. A set of test cases that are matter of interest in the industrial domain are detailed and discussed. Moreover, some insights on the choices done to satisfy the different constraints imposed by the industrial domain are provided too.

Joint work with Christophe Reboud and Pierre Calmon.

PDF abstract: uqsay12_abstract_rmiorelli.pdf.

Organizers: Julien Bect (L2S), Emmanuel Vazquez (L2S), Didier Clouteau (MSSMAT), Filippo Gatti (MSSMAT), Fernando Lopez Caballero (MSSMAT), Bertrand Iooss (EDF R&D).

Practical details: the seminar will be held online using Microsoft Teams.

If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay group on Teams, simply send an email and you will be invited. Please specify which email address the invitation must be sent to (this has to be the address associated with your Teams account).

You will find the link to the seminar on the "General" UQSay channel on Teams, approximately 15 minutes before the beginning.

The technical side of things: you can use Teams either directly from you web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.

Wednesday, June 17, 2020

UQSay #11


The eleventh UQSay seminar on Uncertainty Quantification and related topics, organized by L2S, MSSMAT, and EDF R&D, will take place online on Thursday afternoon, June 25, 2020.

14h–15h — Guilhem Lavabre (EM2C) — [slides]

Dimension reduction and surrogate-modelling for uncertain auto-igniting flame simulations

Numerical simulations have become a backbone of research and industrial design, but few are as expensive as high-fidelity combustion simulations. These combine unsteady turbulent flow and chemistry computations, meaning a few tens of transport equations on meshes of several million nodes. As a result, such simulations can cost up to several millions of CPU-hours.

Alas, all this computational complexity does not shield the results from uncertainty as many sources of variability remain. To name a few: model calibration, operating conditions, exact composition of complex fuels, geometric variability of combustion chambers…

In this context, very sample-efficient methods are needed to make uncertainty propagation even affordable.

This presentation will focus on proposing methods to propagate chemical and experimental uncertainties in simulations of an auto-igniting flame. A toy-problem subject to similar physical phenomena will be introduced. It will be used to implement some uncertain dimension reduction methods. Different surrogate modelling approaches will then be tested and compared to put forward an affordable and relevant design of experiment for this problem.

Joint work with Olivier Gicquel and Ronan Vicquelin.

PDF abstract with illustrations: uqsay11_abstract_glavabre.pdf.

Organizers: Julien Bect (L2S), Emmanuel Vazquez (L2S), Didier Clouteau (MSSMAT), Filippo Gatti (MSSMAT), Fernando Lopez Caballero (MSSMAT), Bertrand Iooss (EDF R&D).

Practical details: the seminar will be held online using Microsoft Teams.

If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay channel on Teams, simply send an email and you will be invited. You will find the link to the seminar on this channel, approximately 15 minutes before the beginning.

The technical side of things: you can use Teams either directly from you web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.