Topic selection & color key :

Topics

All topics are shown.

Forms of presentation

Forms of presentation

All types are shown.

SIAM Conference on Uncertainty Quantification (UQ20)

Friday – 27.03.2020

08:30

iCal
Andreas Mang | Department of Mathematics, University of Houston | United States

Clemens Heitzinger | Vienna University of Technology | Austria

Linus Seelinger | Institute for Scientific Computing, Heidelberg University | Germany

Paul-Remo Wagner | Chair of Risk, Safety and Uncertainty Quantification, ETH Zuerich | Switzerland

MS081: Uncertainty Quantification for Data-Intensive Inverse Problems and Learning (Part I of II)

Chair(s)
Andreas Mang (University of Houston)

Tan Bui-Thanh (The Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin)

Tan Bui-Thanh (The Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin)

Room:
MW HS 2001

Topic:
Inverse problems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Inverse and big-data problems are widespread in computational sciences and engineering. Despite formidable advances in recent years on all frontiers, ranging from pure mathematics to computational sciences, significant challenges remain, especially when it comes to addressing data-driven problems. In inverse/learning problems, parameters are typically related to indirect measurements by a system of partial differential equations (PDEs) or a network, which could be highly nonlinear and nonconvex. Available indirect data are often noisy, and subject to natural variation, while the unknown parameters of interest are high dimensional, or possibly infinite-dimensional in principle. Bayesian inference provides a systematic framework that rigorously that allows us to quantify the uncertainty in the inverse/learning problems, and to assess model validity and adequacy. Since the amount of data we wish to process is only going to increase for the foreseeable future, there is a critical need for effective algorithms that integrate data with simulations and learning approaches that are computation- and data-scalable. This minisymposium aims to attract researchers at the forefront of inverse and learning problems, data science, and data-intensive problems to present their latest work on computation- and data-scalable algorithms in inverse problems and learning.

08:30

Uncertainty Quantification for Inverse Transport Problems

09:00

Computational Bayesian Inversion for Nanocapacitor-Array Biosensors and Electrical-Impedance Tomography

09:30

Multilevel and Multiindex MCMC for High Performance Computing

10:00

Stochastic spectral embedding as a local surrogate model in MCMC based Bayesian model calibration

08:30

iCal
Andrea Arnold | Worcester Polytechnic Institute | United States

Fabien Raphel | Sorbonne Université | France

Daniela Calvetti | Case Western Reserve University | United States

Kevin Flores | North Carolina State University | United States

MS071: Inverse Problems and Uncertainty Quantification in Biological and Medical Applications (Part I of II)

Chair(s)
Andrea Arnold (Worcester Polytechnic Institute)

Sarah Olson (Worcester Polytechnic Institute)

Sarah Olson (Worcester Polytechnic Institute)

Room:
MW HS 0001

Topic:
Physics-informed and data-driven UQ methods

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Uncertainty plays a major role in using mathematics to address biological and medical questions, specifically when analyzing real-world data. This minisymposium features recent mathematical and computational advances in solving inverse problems and quantifying uncertainties for a wide variety of biological and biomedical applications. Topics include development of numerical methods, model reduction, parameter estimation, and data-driven approaches for applications such as safety pharmacology, cell metabolism, tumor growth, and blood coagulation.

08:30

Quantifying Uncertainty in Time-Varying Parameters for Biological Systems

09:00

UQ and estimation of Quantities of Interest in Safety Pharmacology applications

09:30

- CANCELED - Exploring posterior samples of highly underdetermined models

10:00

Equation learning and uncertainty quantification for biological transport models of glioblastoma growth

08:30

iCal
Nick Dexter | Simon Fraser University | Canada

Guannan Zhang | Oak Ridge National Laboratory | United States

Joe Daws | University of Tennessee | United States

Bosu Choi | University of Texas at Austin | United States

MS601: Deep learning and sparse approximation for high-dimensional problems in uncertainty quantification (Part I of II)

Room:
MW HS 1801

Topic:
High-dimensional approximation

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Over the past two decades, we've witnessed two revolutions in applied mathematics and high-dimensional approximation: the rise of sparse reconstruction techniques driven by compressed sensing, and a transformation in data science driven by machine learning with deep neural networks, a.k.a, deep learning. The former seeks to find a compressible representation of a given target function or signal, exploiting structure such as sparsity, parametric smoothness, or low-dimensionality of the solution manifold. The latter seeks to construct a nonlinear approximation from a given dataset, which generalizes well on unseen data points, through a series of compositions of affine and nonlinear mappings. This minisymposium highlights connections between these two topics, with particular attention to recent advances in the theory and algorithms in both approaches, as applied to problems in uncertainty quantification. By bringing together researchers from these two emerging fields, we hope to foster discussion and collaboration on novel theoretical and computational advances in sparse approximation and deep learning, leading to new directions for research.

08:30

Practical Approximation with Deep ReLU Neural Networks

09:00

- CANCELED - Nonlinear level-set learning for dimensionality reduction in high-dimensional function approximation

09:30

Deep Neural Networks inspired by quasi-optimal polynomial approximations for parameterized PDEs

10:00

Sparse Harmonic Transform : Best s-Term Approximation Guarantees for High-Dimensional Functions in Sublinear-Time

08:30

iCal
Marco Scavino | KAUST | Saudi Arabia

Philipp Trunschke | WIAS Berlin | Germany

Simona Dobrilla | TU Braunschweig | Germany

Sharana Kumar Shivanand | TU Braunschweig | Germany

MS042: Propagation of uncertainties and parameter inference in material science (Part II of III)

Chair(s)
Alexander Litvinenko (RWTH Aachen)

Bojana Rosic (University of Twente)

Sebastian Krumscheid (RWTH Aachen)

Bojana Rosic (University of Twente)

Sebastian Krumscheid (RWTH Aachen)

Room:
MW HS 0350

Topic:
UQ for complex systems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

The MS focuses on the process of modeling, quantifying and estimating the effects of uncertainties that characterize irreversible/dissipative material behavior in quasi-static and dynamic conditions. Particular examples of great significance include metal fatigue and concrete fracture analysis, as well as material aging of bone tissues. Moreover, special attention will be paid to the multi-scale and multi-fidelity nature of these problems, as well as to Bayesian analysis and corresponding design of experiments. Numerical tools to be discussed are low-rank functional approximations, Bayesian learning, optimization, stochastic Galerkin, polynomial chaos expansion, and stochastic homogenization, to name just a few.

08:30

Bayesian study of metallic fatigue data considering spatial Poisson processes

09:00

A weighted least squares approach for Bayesian inversion

09:30

- CANCELED - Uncertainty Propagation in a reinforced concrete damage model with localised failure

10:00

Stochastic material modelling and multifidelity uncertainty quantification of bone tissue

08:30

iCal
Kathleen Schmidt | Lawrence Livermore National Laboratory | United States

Jesse Adams | Mission Support and Test Services | United States

Jayaraman J. Thiagarajan | Lawrence Livermore National Laboratory | United States

MS481: UQ at the US DOE National Labs (Part I of II)

Chair(s)
Kathleen Schmidt (Lawrence Livermore National Laboratory)

Kevin Quinlan (Lawrence Livermore National Laboratory)

Kevin Quinlan (Lawrence Livermore National Laboratory)

Room:
MW HS 0250

Topic:
Statistical methods for UQ

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The United States Department of Energy (DOE) Laboratory System grew out of the federally-funded scientific developments of World War II. Today, the national laboratories comprise one of the world’s largest scientific research systems. Tackling areas such as environmental modeling, precision medicine, and global security, the DOE laboratories are at the forefront of scientific innovation and, thus, have access to unique research problems, data sets, and facilities. This minisymposium will showcase the many applications and innovations in UQ stemming from the challenges of the national lab environment.

LLNL-ABS-791303. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

08:30

Data Compositing: Aligning Signals from Asynchronous Sources

09:00

A Gibbs Blocking Scheme for Large-scale Deconvolution

10:00

Building Calibrated Deep Models via Uncertainty Matching with Auxiliary Interval Predictors

08:30

iCal
Ralph Smith | North Carolina State University | United States

Xujia Zhu | ETH Zurich | Switzerland

Alexis Cousin | IFPEN and Ecole Polytechnique | France

Gildas Mazo | INRA | France

MS151: Uncertainty Quantification and Surrogate Models for Stochastic Simulators (Part I of II)

Room:
MW HS 1250

Topic:
Surrogate models

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

Computer simulation models a.k.a. simulators are used nowadays in virtually all fields of applied science and engineering. Usually, simulators that predict quantities of interests (QoI) as a function of input parameters are deterministic, i.e. they can be considered as a mapping from an input- to an output space. Running the simulator twice with the same input values provides identical outputs.

In contrast, so-called stochastic simulators contain hidden sources of uncertainty (e.g. latent variables) or uncontrollable inputs, on top of the well-identified and controllable inputs, meaning that repeated runs with the same inputs provides different results. Of interest is the resulting distribution of the QoI conditioned by the input (controllable) parameters. This distribution can be characterized in a rough way by replicating the runs of the simulator for the same controllable inputs. Unfortunately, in the context of uncertainty propagation and sensitivity analysis, handling stochastic simulators may be highly demanding due to these replications. One appropriate solution can be to use surrogate models (also referred to metamodels) to approximate the conditional expectation of the model, from a limited number of simulations.

In this MS, we will present recent developments in the field of surrogate models for stochastic simulators, be it for uncertainty propagation, sensitivity analysis or robust design.

08:30

Surrogate Model Development for Radiation Source Localization Using 3-D monte Carlo Transport Codes

09:00

Sparse generalized lambda models for surrogating stochastic simulators

09:30

Chance constraint optimization of a complex system - Application to the design of a floating offshore wind turbine

10:00

Tradeoffs between explorations and repetitions in variance-based sensitivity analysis for stochastic models

08:30

iCal
Bernd Heidergott | Vrijie Universiteit Amsterdam | Netherlands

Jeremiah Birrell | University of Massachusetts Amherst | United States

Sung-Ha Hwang | Korea Advanced Institute of Science and Technology | Korea, Republic of

Henry Lam | Columbia University | United States

MS521: Model Uncertainty, Robust Optimization and Predictive Guarantees (Part I of II)

Chair(s)
Luc Rey-Bellet (University of Massachusetts Amherst)

Markos Katsoulakis (University of Massachusetts Amherst)

Henry Lam (Columbia University)

Markos Katsoulakis (University of Massachusetts Amherst)

Henry Lam (Columbia University)

Room:
MW 0608m

Topic:
Probability Theory for UQ

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The main topics of the mini-symposium include model uncertainty, robust uncertainty quantification & optimization, and their implications in predictive modeling guarantees and rare-event analysis. We aim at bringing together closely related but possibly disparate communities in applied mathematics, applied probability, information theory, operations research, optimization and economics, to foster interdisciplinary discussions and collaborations. Speakers will demonstrate recent mathematical and conceptual developments of related UQ methods, and also their applications ranging from engineering design of materials to econometrics and risk analysis.

08:30

Input vs. Output Modelling: Towards Measuring the Accuracy of Models

09:00

Information-Theoretic Approaches to Distributional Robustness

09:30

- CANCELED - Ambiguity aversion in Incomplete Information Games: Auctions and Oligopolistic competition

10:00

Computationally Efficient Quantification of Simulation Input Uncertainty

08:30

iCal
Harri Hakula | Aalto University | Finland

Troy Butler | University of Colorado Denver | United States

Lukas Bruder | Technical University of Munich | Germany

Tim Wildey | Sandia National Labs | United States

MS831: Stories of Marrying Methods and Applications (Part I of II)

Room:
MW ZS 1050

Topic:
UQ for complex systems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

It is a story as old as time. Models rife with uncertainty are developed for intriguing applications while simultaneously uncertainty quantification (UQ) methods are rapidly advanced. Yet, when the developers of the models and methods meet, it is rarely love at first sight. Either the UQ questions the modeler asks are like the third cousin to those the methods are intended to answer or the methods require certain types or quantities of data for which the modeler is not prepared to deliver. This minisymposium brings together pairs of collaborative researchers giving coordinated presentations on how an application and UQ method were finally joined in harmony. The first presentation focuses on the application, modeling, and types of UQ questions the researchers seek to answer. The second presentation focuses on how a UQ method was tailored to answer these questions under the constraints of the model.

08:30

Striking the Right Chord (Part I): What note is this?

09:00

Striking the Right Chord (Part II): Learning to hear

09:30

Data-consistent computational modeling of the mechanical behavior of abdominal aortic aneurysms

10:00

A mathematical and computational framework for developing data-consistent probability densities with application to abdominal aortic aneurysms

08:30

iCal
Jean-Michel Brankart | Universite Grenoble Alpes | France

Boujemaa Ait-El-Fquih | King Abdullah University of Science and Technology (KAUST) | Saudi Arabia

Yue (Michael) Ying | National Center for Atmospheric Research | United States

MS371: Data Assimilation: Methods and Applications Earth System Models (Part I of II)

Chair(s)
Ian Grooms (University of Colorado Boulder)

Aneesh Subramanian (University of Colorado Boulder)

Ibrahim Hoteit (KAUST)

Aneesh Subramanian (University of Colorado Boulder)

Ibrahim Hoteit (KAUST)

Room:
MW ZS 2050

Topic:
Data assimilation

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Data assimilation in Earth system models combines high-dimensional, coupled, nonlinear models with large volumes of in situ and remotely sensed observational data. The dynamics and observations are nonlinear, the distributions are non-Gaussian, and the cost of simulation is high. The goal of the minisymposium is to provide a forum for this diverse group to discuss and share ideas for advancing the science of DA in climate modeling or any of its components (e.g. atmosphere, ocean, ice sheets, land models, or sea ice). Topics of interest include coupled data assimilation; strategies for estimating and mitigating model errors; strategies for addressing strong nonlinearities and non-Gaussianity; multiscale, multilevel, or multifidelity methods; and machine learning methods for data assimilation.

08:30

Implicitly localized MCMC sampler to cope with nonlocal/nonlinear data constraints in large-size inverse problems

09:30

- CANCELED - Parametric Bayesian estimation of point-like pollution sources of groundwater layers

10:00

- CANCELED - Developing data assimilation algorithms for the analysis and prediction of geophysical flows across many scales

08:30

iCal
Armenak Petrosyan | Oak Ridge National Lab | United States

Vishagan Ratnaswamy | Sandia National Lab | United States

Jeremiah Hauth | University of Michigan Ann Arbor | United States

Liu Yang | Brown University | United States

MS401: Recent Advances in Data-driven Modeling for Uncertainty Quantification (Part I of II)

Chair(s)
Jiahua Jiang (Virginia Tech)

Ling Guo (Shanghai Normal University)

Jing Li (Pacific Northwest National Laboratory)

Ling Guo (Shanghai Normal University)

Jing Li (Pacific Northwest National Laboratory)

Room:
MW ZS 1450

Topic:
Physics-informed and data-driven UQ methods

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The amount of data in existence is growing exponentially. This has lead to the development of an unavoidable basin of attraction in data-driven uncertainty quantification (UQ) approaches for large-scale or high dimensional (UQ) problems. However, it is still in its infancy and new ideas are needed for this core research area.

The goal of our minisymposium is to provide a forum for this diverse group to discuss and share ideas for developing data-driven UQ approaches. These advanced UQ methods involve (but are not limited to) machine learning, neural network, model reduction as well as advances in Bayesian framework. Various applications will be used to show the performance of these improved UQ approaches.

08:30

- CANCELED - Integral Representations for Neural Networks with Applications to Uncertainty Quantification

09:00

Physics-Informed Recurrent Neural Networks for Land Model Surrogate Construction for Uncertainty Quantification

09:30

Capturing Training Uncertainty Using Bayesian Neural Networks in a Rotorcraft Ice Detection Application

10:00

Uncertainty Quantification for Nonlinear PDEs with Noisy Data Using Deep Flow-based Generative Models

08:30

iCal
Kevin Carlberg | Sandia National Laboratories Livermore | United States

Liu Liu | The University of Texas at Austin | United States

Yanyan He | University of North Texas | United States

Mariella Kast | Ecole polytechnique fédérale de Lausanne | Switzerland

MS881: Recent Advances in Machine Learning and Data-driven Methods for Modeling Uncertainty in Computational Science and Engineering (Part I of II)

Chair(s)
Xueyu Zhu (University of Iowa)

Akil Narayan (University of Utah)

Yanyan He (University of North Texas)

Akil Narayan (University of Utah)

Yanyan He (University of North Texas)

Room:
MW ZS 1550

Topic:
Physics-informed and data-driven UQ methods

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Despite the remarkable growth in computational power, it is still very computationally expensive to simulate most real-world systems in full detail, including a comprehensive analysis of parameter and model uncertainty. In such situations, data-driven approaches are tractable computational methods that provide useful empirical characterizations of uncertainty and have been successfully exploited in recent years. The increasing availability of very large data sets for this purpose makes techniques in machine learning an attractive toolbox for uncertainty emulation and characterization.

This minisymposium focuses on recent advances in uncertainty quantification algorithmic developments and applications based on data-driven and machine learning approaches in large-scale applications. Topics include data-driven surrogate construction, data assimilation, and physics-informed machine learning based on a limited number of data/observations and provide guidance for the system design, forecasting, etc.

08:30

- CANCELED - Model reduction of dynamical systems with deep convolutional autoencoders: hyper-reduction, ResNets, and application to UQ

09:00

- CANCELED - A bi-fidelity stochastic collocation method for multiscale kinetic equations with random parameters

09:30

- CANCELED - Optimal allocation of computational resources based on Gaussian Process

10:00

Sensor data integration using physics-informed, boundary-aware Gaussian processes

08:30

iCal
Jerrad Hampton | University of Colorado Boulder | United States

Liang Yan | Southeast University | China

Terrence Alsup | New York University | United States

Simon Cotter | University of Manchester | United Kingdom

MS813: Multilevel and Multi-fidelity Methods for Model-Based Statistical Learning (Part III of III)

Chair(s)
Tiangang Cui (Monash University)

Santiago Badia (Monash University)

Alexander Gilbert (Ruprecht-Karls University Heidelberg)

Youssef Marzouk (Massachusetts Institute of Technology)

Robert Scheichl (Ruprecht-Karls University Heidelberg)

Santiago Badia (Monash University)

Alexander Gilbert (Ruprecht-Karls University Heidelberg)

Youssef Marzouk (Massachusetts Institute of Technology)

Robert Scheichl (Ruprecht-Karls University Heidelberg)

Room:
MW 2250

Topic:
Statistical methods for UQ

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

The synthesis of various information sources, including a priori domain knowledge, statistical assumptions, field data, etc., large-scale numerical models is one of the key steps in building interpretable and predictive models for supporting critical decisions in science, engineering, medicine, and beyond. Typical examples can be found in oil/gas reservoir modeling, treatment of saltwater intrusion, medical imaging, tumor treatment, aircraft design. Because of the computationally costly nature of the numerical models and stringent requirements on the accuracy of the statistical learning outcomes, multilevel and multi-fidelity methods provide a viable route for solving these model-based statistical learning tasks. This mini-symposium will bring together researchers working on the forefront of multilevel and multi-fidelity methods (and other relevant methods) intended to accelerate model-based statistical learning tasks.

08:30

Embedded multilevel Monte Carlo for UQ on random domains

09:00

- CANCELED - Adaptive Multi-fidelity surrogate modeling for Bayesian inference in inverse problems

09:30

Context-aware model reduction for multifidelity importance sampling

10:00

Squashing the banana: Transport map-accelerated adaptive importance sampling

08:30

iCal
Mattia de' Michieli Vitturi | Istituto Nazionale di Geofisica e Vulcanologia (INGV) | Italy

Riccardo Pellegrini | Istituto di Ingegneria del Mare (CNR-INM) | Italy

Saleh Rezaeiravesh | Royal Institute of Technology (KTH) | Sweden

Maria Vittoria Salvetti | University of Pisa | Italy

MS511: UQ for complex fluid dynamics problems in realistic applications (Part I of II)

Chair(s)
Lorenzo Tamellini (Istituto di Matematica Applicata e Tecnologie Informatiche (CNR-IMATI))

Matteo Diez (Istituto di Ingegneria del Mare (CNR-INM))

Maria Vittoria Salvetti (University of Pisa)

Matteo Diez (Istituto di Ingegneria del Mare (CNR-INM))

Maria Vittoria Salvetti (University of Pisa)

Room:
MW 1701

Topic:
UQ for complex systems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

Uncertainty Quantification techniques are by now mature enough to address realistic, large scale problems of significant relevance. In this minisymposium, we focus in particular on complex fluid dynamics problems for engineering and environmental applications, that are fields in which computational science has traditionally played a major role. Several different kinds of UQ analyses naturally arise in these fields: forward UQ, optimization under uncertainty, inverse problem and data assimilation (e.g. for real-time control). In these scenarios, non-standard randomness might occur, and the complexity of the governing PDE equations further introduces significant and fascinating theoretical and computational challenges. In particular, polynomial-based UQ methods like Polynomial Chaos or Sparse grids collocation might not work well, in which case one has to resort to sampling methods, control variates, and more recently, machine learning techniques. Ad-hoc algorithms for high-performance computing are also relevant in this framework and welcome in this minisymposium.

08:30

Uncertainty quantification in numerical modeling of volcanic ash dispersal in the atmosphere

09:00

- NEW - Adaptive Multi-fidelity Surrogates for Uncertainty Quantification of Noisy CFD Data

09:30

On Numerical and Statistical Uncertainties in Scale-Resolving Simulations of Wall Turbulence

10:00

Impact of uncertainties in inlet conditions in numerical simulations of the blood flow in the thoracic aorta

08:30

iCal
Whitney Huang | Clemson University | United States

Pulong Ma | Statistical and Applied Mathematical Sciences Institute | United States

Serge Guillas | University College London | United Kingdom

Devaraj Gopinathan | University College London | United Kingdom

MS271: The Science of Hazards: Tsunami and Storm Surges

Chair(s)
E. Bruce Pitman (University at Buffalo)

Serge Guillas (University College London and Turing Institute)

Whitney Huang (Clemson University)

Serge Guillas (University College London and Turing Institute)

Whitney Huang (Clemson University)

Room:
IAS 0.001

Topic:
Prediction

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Computer modeling makes possible the simulation of shoreline hazards, from tsunamis and tropical storms. To make predictions of these hazard events requires many simulations, to explore the high dimensional space of input parameters, and massive computational budgets. Data science methods are needed to detect nascent storms and tsunami waves and feed information to simulation models, to monitor the evolving hazard or make long-term predictions. Statistical emulators can estimate the output of simulations and greatly reduce the computational burden. However the necessary outputs are often spatio-temporal fields, and conventional methods for constructing emulators cannot be applied. This mini-symposium, which emerged from research activity during the 2018-19 SAMSI program on Uncertainty Quantification, will bring together scientists working on computational and statistical methodology to better predict and track storms and tsunamis.

08:30

A Combined Physical-Statistical Approach for Estimating Storm Surge Risk

09:00

Uncertainty Quantification in Assessing Storm Surge Hazards

09:30

How high could be the coastal waves generated by a storm surge? A computationally efficient surrogate-based optimization approach

10:00

Uncertainty Quantification of Tsunami Currents and Heights

08:30

iCal
Christian P. Robert | Universite Paris-Dauphine | France

Florence Forbes | Universite Grenoble Alpes, Inria, CNRS | France

Matti Vihola | University of Jyvaskyla | Finland

Ann Lee | Carnegie Mellon University | United States

MS751: Advances in likelihood-free inference (Part I of II)

Chair(s)
Ricardo Baptista (Massachusetts Institute of Technology)

Giulio Trigila (Baruch College of New York)

Giulio Trigila (Baruch College of New York)

Room:
IAS 4.001

Topic:
Inverse problems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Many statistical models of interest in engineering, the sciences, and machine learning define a likelihood function that is computationally prohibitive to evaluate. This may be induced from the model only being known through a data generating process or the likelihood function involving a high-dimensional integral (e.g., from a marginalization procedure or the computation of a normalizing constant). In these cases, it is difficult to apply classical inference methods such as maximum likelihood estimation or likelihood-based Bayesian inference algorithms. To enable inference in these settings, several approaches have been developed in the statistics and machine learning community that avoid direct evaluation of the likelihood function (e.g., approximate Bayesian computation). Despite these success, efficiently solving such problems remains challenging, especially in high dimensions, or when only limited information or few samples are available. This mini-symposium will explore new algorithms and methodologies for performing likelihood-free inference in these complex models.

08:30

Component-wise approximate Bayesian computation via Gibbs-like steps

09:00

Approximate Bayesian computation via the energy statistic

09:30

On the use of approximate Bayesian computation Markov chain Monte Carlo with inflated tolerance and post-correction

10:00

Confidence Regions and Hypothesis Testing in a Likelihood-Free Inference Setting

08:30

iCal
Arnaud Guyader | Sorbonne Université | France

Fabian Wagner | Technical University of Munich | Germany

Tengchao Yu | Shanghai Jiao Tong University | China

Max Ehre | Technical University of Munich | Germany

MS361: Theory and simulation of failure probabilities and rare events (Part I of II)

Chair(s)
Iason Papaioannou (Technical University of Munich)

Michael D. Shields (Johns Hopkins University)

Dimitrios Giovanis (Johns Hopkins University)

Elisabeth Ullmann (Technical University of Munich)

Michael D. Shields (Johns Hopkins University)

Dimitrios Giovanis (Johns Hopkins University)

Elisabeth Ullmann (Technical University of Munich)

Room:
Interims Lecture Hall 101

Topic:
Rare events and Risk

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

The evaluation of failure probabilities is a fundamental problem in reliability analysis and risk management of systems with uncertain inputs. We consider systems described by PDEs with random coefficients together with efficient approximation schemes. This includes stochastic finite elements, collocation, reduced basis, and advanced Monte Carlo methods. Efficient evaluation and updating of small failure probabilities and rare events remains a significant computational challenge. This mini-symposium brings together tools from applied probability, numerical analysis, and computational science and engineering. We showcase advances in analysis and computational treatment of rare events and failure probabilities, including variance reduction, advanced meta-models, and multilevel Monte Carlo.

08:30

On the Asymptotic Normality of Adaptive Multilevel Splitting

09:00

Error analysis of probabilities of rare events with approximate models

09:30

A Weight-bounded Importance Sampling Method for Variance Reduction

10:00

Conditional reliability estimation with importance sampling based on information reuse

08:30

iCal
Sahani Pathiraja | University of Potsdam | Germany

Svetlana Dubinkina | Centrum Wiskunde & Informatica (CWI) | Netherlands

MS771: Ensemble & Particle methods for inverse problems (Part I of II)

Room:
Interims Lecture Hall 102

Topic:
Inverse problems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The understanding and incorporation of data within models has become a vital component of applied mathematics. A fundamental one can ask is given noisy measurements of data, how to recover some unknown quantity of interest. Some examples of these fields include in- verse problems which is primarily concerned with parameter estimation and data assimilation for state estimation. Both fields have seen a considerable amount of attention due to recent advance- ments in terms of both classical and statistical approaches. In particular, this mini-symposium will consider particle methods for solving inverse problems with the help optimization tools as well as particle methods aiming to represent the posterior distribution in a bayesian point of view for inverse problems.

The motivation behind this mini-symposium is to bring together experts from both schools. This would provide a complimentary field to the mini-symposium where connections between both areas are currently being developed.

08:30

Discrete Gradient methods for Computational Bayesian inverse problems

09:00

Bayesian approach to elliptic inverse problems

08:30

iCal
Xun Huan | University of Michigan | United States

Antony Overstall | University of Southampton | United Kingdom

Kareem Abdelfatah | University of South Carolina | United States

Markus Hainy | Johannes Kepler University | Austria

MS891: Model-based Optimal Experimental Design (Part I of II)

Room:
Exzellenzzentrum 0003

Topic:
Design of experiments

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The challenge of acquiring the most valuable data from experiments—for the purpose of inference, prediction, design, or control—has received substantial attention in statistics, applied mathematics, and engineering and science. This task can be formalized through the framework of optimal experimental design (OED). Models describing experimental conditions and processes, both physical and statistical, can be particularly useful for arriving at these optimal designs. However, model-based OED faces many challenges, such as formulational difficulties, choices of optimality criteria, computation of information metrics, handling nonlinear responses and non-Gaussian distributions, and dealing with expensive and dynamically evolving simulations. This minisymposium invites researchers of model-based optimal experimental design, in the broad areas of computational and applications-oriented developments.

08:30

Optimal Bayesian Design of Sequential Experiments Using Reinforcement Learning with Policy Gradient Method

09:00

Bayesian design of experiments for an alternative model

09:30

- CANCELED - Active Learning in Computational Catalysis using Stacked Gaussian Processes

10:00

Optimal Bayesian design for models with intractable likelihoods via supervised learning methods

11:00

iCal
Karen Veroy-Grepl | Eindhoven University of Technology (TU/e) | Netherlands

IP06: Karen Veroy-Grepl: Optimal experimental design for the quantification of model uncertainty: A functional analysis perspective

Room:
MW HS 2001

Topic:
Design of experiments

Form of presentation:
Plenary Lecture

Duration:
45 Minutes

11:00

Optimal experimental design for the quantification of model uncertainty: A functional analysis perspective

11:00

iCal

IP06 - streamed from HS 2001: Karen Veroy-Grepl: Optimal experimental design for the quantification of model uncertainty: A functional analysis perspective

Room:
MW HS 0001

Topic:
Design of experiments

Form of presentation:
Plenary Lecture

Duration:
45 Minutes

11:45

iCal
Youssef Marzouk | Massachusetts Institute of Technology | United States

IP07: Youssef M. Marzouk: Transport methods for stochastic modeling and inference

Room:
MW HS 2001

Topic:
Statistical methods for UQ

Form of presentation:
Plenary Lecture

Duration:
45 Minutes

11:45

Transport methods for stochastic modeling and inference

11:45

iCal

IP07 - streamed from HS 2001: Youssef M. Marzouk: Transport methods for stochastic modeling and inference

Room:
MW HS 0001

Topic:
Statistical methods for UQ

Form of presentation:
Plenary Lecture

Duration:
45 Minutes

14:00

iCal
Dongbin Xiu | Ohio State University | United States

Xiao Chen | Lawrence Livermore National Laboratory | United States

Sheroze Sheriffdeen | Oden Institute of Computational Engineering and Sciences, The University of Texas at Austin | United States

MS082: Uncertainty Quantification for Data-Intensive Inverse Problems and Learning (Part II of II)

Chair(s)
Andreas Mang (University of Houston)

Tan Bui-Thanh (The Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin)

Tan Bui-Thanh (The Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin)

Room:
MW HS 2001

Topic:
Inverse problems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Inverse and big-data problems are widespread in computational sciences and engineering. Despite formidable advances in recent years on all frontiers, ranging from pure mathematics to computational sciences, significant challenges remain, especially when it comes to addressing data-driven problems. In inverse/learning problems, parameters are typically related to indirect measurements by a system of partial differential equations (PDEs) or a network, which could be highly nonlinear and non-convex. Available indirect data are often noisy, and subject to natural variation, while the unknown parameters of interest are high dimensional, or possibly infinite-dimensional in principle. Bayesian inference provides a systematic framework that rigorously that allows us to quantify the uncertainty in the inverse/learning problems, and to assess model validity and adequacy. Since the amount of data we wish to process is only going to increase for the foreseeable future, there is a critical need for effective algorithms that integrate data with simulations and learning approaches that are computation- and data-scalable. This minisymposium aims to attract researchers at the forefront of inverse and learning problems, data science, and data-intensive problems to present their latest work on computation- and data-scalable algorithms in inverse problems and learning.

14:30

Data Driven Governing Equations Recovery with Deep Neural Networks

15:00

- NEW - A Bayesian Approach to Real-Time Dynamic Parameter Estimation Using PMU Measurement

15:30

Bayesian Inverse Problems Using Dimensionality Reduction and Machine Learning

14:00

iCal
Anastasios Matzavinos | Brown University | United States

Erkki Somersalo | Case Western Reserve University | United States

Laura Albrecht | Colorado School of Mines | United States

Nick Cogan | Florida State University | United States

MS072: Inverse Problems and Uncertainty Quantification in Biological and Medical Applications (Part II of II)

Chair(s)
Andrea Arnold (Worcester Polytechnic Institute)

Sarah Olson (Worcester Polytechnic Institute)

Sarah Olson (Worcester Polytechnic Institute)

Room:
MW HS 0001

Topic:
Physics-informed and data-driven UQ methods

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Uncertainty plays a major role in using mathematics to address biological and medical questions, specifically when analyzing real-world data. This minisymposium features recent mathematical and computational advances in solving inverse problems and quantifying uncertainties for a wide variety of biological and biomedical applications. Topics include development of numerical methods, model reduction, parameter estimation, and data-driven approaches for applications such as safety pharmacology, cell metabolism, tumor growth, and blood coagulation.

14:00

Bayesian uncertainty quantification for particle-based simulation of lipid bilayer membranes

14:30

- CANCELED - Model uncertainties in gas transport to cells

15:00

Experimental design and identifiability in models of blood coagulation

15:30

Contrasting Sensitivity and Data Assimilation in the Context of the Autoimmune Disease Alopecia Areata

14:00

iCal
Simone Brugiapaglia | Concordia University | Canada

Philipp Petersen | University of Vienna | Austria

Mones Raslan | Technical University of Berlin | Germany

Timo Welti | ETH Zurich | Switzerland

MS602: Deep learning and sparse approximation for high-dimensional problems in uncertainty quantification (Part II of II)

Room:
MW HS 1801

Topic:
High-dimensional approximation

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Over the past two decades, we've witnessed two revolutions in applied mathematics and high-dimensional approximation: the rise of sparse reconstruction techniques driven by compressed sensing, and a transformation in data science driven by machine learning with deep neural networks, a.k.a, deep learning. The former seeks to find a compressible representation of a given target function or signal, exploiting structure such as sparsity, parametric smoothness, or low-dimensionality of the solution manifold. The latter seeks to construct a nonlinear approximation from a given dataset, which generalizes well on unseen data points, through a series of compositions of affine and nonlinear mappings. This minisymposium highlights connections between these two topics, with particular attention to recent advances in the theory and algorithms in both approaches, as applied to problems in uncertainty quantification. By bringing together researchers from these two emerging fields, we hope to foster discussion and collaboration on novel theoretical and computational advances in sparse approximation and deep learning, leading to new directions for research.

14:00

Greedy algorithms for sparse high-dimensional function approximation

14:30

Deep ReLU neural networks and finite element spaces

15:00

Solving Parametric PDEs with Deep Neural Networks: A Theoretical and Numerical Analysis

15:30

- NEW - Full strong error analysis for the training of deep neural networks with stochastic gradient descent

14:00

iCal
Giacomo Sevieri | University College London | United Kingdom

Dieter Moser | RWTH Aachen | Germany

Dmitry Kabanov | RWTH Aachen University | Germany

Nadhir Ben Rached | RWTH Aachen University | Germany

MS043: Propagation of uncertainties and parameter inference in material science (Part III of III)

Chair(s)
Alexander Litvinenko (RWTH Aachen)

Bojana Rosic (University of Twente)

Sebastian Krumscheid (RWTH Aachen)

Bojana Rosic (University of Twente)

Sebastian Krumscheid (RWTH Aachen)

Room:
MW HS 0350

Topic:
UQ for complex systems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

The MS focuses on the process of modeling, quantifying and estimating the effects of uncertainties that characterize irreversible/dissipative material behavior in quasi-static and dynamic conditions. Particular examples of great significance include metal fatigue and concrete fracture analysis, as well as material aging of bone tissues. Moreover, special attention will be paid to the multi-scale and multi-fidelity nature of these problems, as well as to Bayesian analysis and corresponding design of experiments. Numerical tools to be discussed are low-rank functional approximations, Bayesian learning, optimization, stochastic Galerkin, polynomial chaos expansion, and stochastic homogenization, to name just a few.

14:00

The effect of epistemic uncertainties on the seismic fragility assessment of existing concrete gravity dams

14:30

A low-rank surogate for a gradient-extended damage-plasticity model

15:00

Physics-informed deep learning of divergence-free flows

15:30

Importance Sampling for a Robust and Efficient Multilevel Monte Carlo Estimator for Stochastic Biological Systems

14:00

iCal
Kevin Quinlan | Lawrence Livermore National Laboratory | United States

James R. Gattiker | Los Alamos National Laboratory | United States

George Ostrouchov | Oak Ridge National Laboratory and University of Tennessee | United States

Adah Zhang | Sandia National Laboratories | United States

MS482: UQ at the US DOE National Labs (Part II of II)

Chair(s)
Kathleen Schmidt (Lawrence Livermore National Laboratory)

Kevin Quinlan (Lawrence Livermore National Laboratory)

Kevin Quinlan (Lawrence Livermore National Laboratory)

Room:
MW HS 0250

Topic:
Statistical methods for UQ

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The United States Department of Energy (DOE) Laboratory System grew out of the federally-funded scientific developments of World War II. Today, the national laboratories comprise one of the world’s largest scientific research systems. Tackling areas such as environmental modeling, precision medicine, and global security, the DOE laboratories are at the forefront of scientific innovation and, thus, have access to unique research problems, data sets, and facilities. This minisymposium will showcase the many applications and innovations in UQ stemming from the challenges of the national lab environment.

LLNL-ABS-791303. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

14:00

Design of Experiments for Agent Based Models

14:30

- CANCELED - Boosting Carbon Capture with Sequential Design

15:00

Scaling and Uncertainty in Model-Based Clustering of Medical Trajectories

15:30

Implementing Bayesian UQ for Complex Systems

14:00

iCal
Alen Alexanderian | North Carolina State University | United States

Thierry Klein | Ecole Nationale de l'Aviation Civile / Université de Toulouse | France

Pierre Etoré | Université Grenoble Alpes | France

MS152: Uncertainty Quantification and Surrogate Models for Stochastic Simulators (Part II of II)

Room:
MW HS 1250

Topic:
Surrogate models

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

Computer simulation models a.k.a. simulators are used nowadays in virtually all fields of applied science and engineering. Usually, simulators that predict quantities of interests (QoI) as a function of input parameters are deterministic, i.e. they can be considered as a mapping from an input- to an output space. Running the simulator twice with the same input values provides identical outputs.

In contrast, so-called stochastic simulators contain hidden sources of uncertainty (e.g. latent variables) or uncontrollable inputs, on top of the well-identified and controllable inputs, meaning that repeated runs with the same inputs provides different results. Of interest is the resulting distribution of the QoI conditioned by the input (controllable) parameters. This distribution can be characterized in a rough way by replicating the runs of the simulator for the same controllable inputs. Unfortunately, in the context of uncertainty propagation and sensitivity analysis, handling stochastic simulators may be highly demanding due to these replications. One appropriate solution can be to use surrogate models (also referred to metamodels) to approximate the conditional expectation of the model, from a limited number of simulations.

In this MS, we will present recent developments in the field of surrogate models for stochastic simulators, be it for uncertainty propagation, sensitivity analysis or robust design.

14:00

Multiscale global sensitivity analysis for stochastic reaction networks

15:00

Stochastic computer codes and sensitivity analysis

15:30

Global sensitivity analysis for models described by stochastic differential equations

14:00

iCal
Eric Hall | RWTH Aachen University | Germany

Guo-Jhen Wu | KTH Stockholm | Sweden

Petr Plechac | University of Delaware | United States

Yixiang Mao | Harvard University | United States

MS522: Model Uncertainty, Robust Optimization and Predictive Guarantees (Part II of II)

Chair(s)
Luc Rey-Bellet (University of Massachusetts Amherst)

Markos Katsoulakis (University of Massachusetts Amherst)

Henry Lam (Columbia University)

Markos Katsoulakis (University of Massachusetts Amherst)

Henry Lam (Columbia University)

Room:
MW 0608m

Topic:
Probability Theory for UQ

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The main topics of the mini-symposium include model uncertainty, robust uncertainty quantification & optimization, and their implications in predictive modeling guarantees and rare-event analysis. We aim at bringing together closely related but possibly disparate communities in applied mathematics, applied probability, information theory, operations research, optimization and economics, to foster interdisciplinary discussions and collaborations. Speakers will demonstrate recent mathematical and conceptual developments of related UQ methods, and also their applications ranging from engineering design of materials to econometrics and risk analysis.

14:00

Predictive Probabilistic Graphical Models for Energy Materials

14:30

Large deviation properties of the empirical measure of a dynamical system subject to small random perturbations

15:00

Likelihood ratio methods for estimating sensitivity and linear response in stochastic dynamics

15:30

Uncertainty quantification for non-absolutely continuous perturbations of probability measures

14:00

iCal
Nuutti Hyvönen | Aalto University | Finland

Juha-Pekka Puska | Aalto University | Finland

Martin Simon | Deka GmbH | Germany

Lassi Roininen | Lappeenranta-Lahti University of Technology | Finland

MS832: Stories of Marrying Methods and Applications (Part II of II)

Room:
MW ZS 1050

Topic:
UQ for complex systems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

It is a story as old as time. Models rife with uncertainty are developed for intriguing applications while simultaneously uncertainty quantification (UQ) methods are rapidly advanced. Yet, when the developers of the models and methods meet, it is rarely love at first sight. Either the UQ questions the modeler asks are like the third cousin to those the methods are intended to answer or the methods require certain types or quantities of data for which the modeler is not prepared to deliver. This minisymposium brings together pairs of collaborative researchers giving coordinated presentations on how an application and UQ method were finally joined in harmony. The first presentation focuses on the application, modeling, and types of UQ questions the researchers seek to answer. The second presentation focuses on how a UQ method was tailored to answer these questions under the constraints of the model.

14:00

Computational framework for applying electrical impedance tomography to head imaging

14:30

Applying approximation error modelling to head imaging by electrical impedance tomography

15:00

Stock Price Bubbles - A Data-Driven Indicator: Practitioner's view

15:30

Stock Price Bubbles - A Data-Driven Indicator: UQ view

14:00

iCal
Ibrahim Hoteit | KAUST | Saudi Arabia

Ian Grooms | University of Colorado Boulder | United States

Aneesh Subramanian | University of Colorado Boulder | United States

MS372: Data Assimilation: Methods and Applications Earth System Models (Part II of II)

Chair(s)
Ibrahim Hoteit (KAUST)

Aneesh Subramanian (University of Colorado Boulder)

Ian Grooms (University of Colorado Boulder)

Aneesh Subramanian (University of Colorado Boulder)

Ian Grooms (University of Colorado Boulder)

Room:
MW ZS 2050

Topic:
Data assimilation

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Data assimilation in Earth system models combines high-dimensional, coupled, nonlinear models with large volumes of in situ and remotely sensed observational data. The dynamics and observations are nonlinear, the distributions are non-Gaussian, and the cost of simulation is high. The goal of the minisymposium is to provide a forum for this diverse group to discuss and share ideas for advancing the science of DA in climate modeling or any of its components (e.g. atmosphere, ocean, ice sheets, land models, or sea ice). Possible topics of interest include coupled data assimilation; strategies for estimating and mitigating model errors; strategies for addressing strong nonlinearities and non-Gaussianity; multiscale, multilevel, or multifidelity methods; and machine learning methods for data assimilation.

14:30

- CANCELED - A Particle Filter-based Adaptive Inflation Scheme for the Ensemble Kalman Filter

15:00

A hybridized EnKF and smoothed-observations particle filter

15:30

- CANCELED - Impact of ocean observation systems on ocean analyses and subseasonal forecasts

14:00

iCal
Yukiko Shimizu | Sandia National Laboratories | United States

Huan Lei | Michigan State University | United States

Xinghui Zhong | Zhejiang University | China

Georgios Karagiannis | Durham University | United Kingdom

MS402: Recent Advances in Data-driven Modeling for Uncertainty Quantification (Part II of II)

Chair(s)
Jiahua Jiang (Virginia Tech)

Ling Guo (Shanghai Normal University)

Jing Li (Pacific Northwest National Laboratory)

Ling Guo (Shanghai Normal University)

Jing Li (Pacific Northwest National Laboratory)

Room:
MW ZS 1450

Topic:
Physics-informed and data-driven UQ methods

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The amount of data in existence is growing exponentially. This has lead to the development of an unavoidable basin of attraction in data-driven uncertainty quantification (UQ) approaches for large-scale or high dimensional (UQ) problems. However, it is still in its infancy and new ideas are needed for this core research area.

The goal of our minisymposium is to provide a forum for this diverse group to discuss and share ideas for developing data-driven UQ approaches. These advanced UQ methods involve in (but not limited to) machine learning, neural network, model reduction as well as advances in Bayesian framework. Various applications will be used to show the performance of these improved UQ approaches.

14:00

Windowed Space-time Least-squares Petrov-Galerkin Projection for Nonlinear Model Reduction

14:30

- CANCELED - Data-driven Approach for Uncertainty Quantification in Complex Systems with Arbitrary Density

15:00

- CANCELED - Uncertainty Quantification for Kinematic Wave Models Based on CDF Methods

15:30

Bayesian UQ analysis of Computer Models with Local Features Under the Presence of Non-nested Multi-fidelity Designs: Application to the WRF Model

14:00

iCal
Takemasa Miyoshi | RIKEN | Japan

Zuoqiang Shi | Tsinghua University | China

Yeonjong Shin | Brown University | United States

Qi Sun | Beijing International Center for Mathematical Research, Peking University | China

MS882: Recent Advances in Machine Learning and Data-driven Methods for Modeling Uncertainty in Computational Science and Engineering (Part II of II)

Chair(s)
Xueyu Zhu (University of Iowa)

Akil Narayan (University of Utah)

Yanyan He (University of North Texas)

Akil Narayan (University of Utah)

Yanyan He (University of North Texas)

Room:
MW ZS 1550

Topic:
Physics-informed and data-driven UQ methods

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Despite the remarkable growth in computational power, it is still very computationally expensive to simulate most real-world systems in full detail, including a comprehensive analysis of parameter and model uncertainty. In such situations, data-driven approaches are tractable computational methods that provide useful empirical characterizations of uncertainty and have been successfully exploited in recent years. The increasing availability of very large data sets for this purpose makes techniques in machine learning an attractive toolbox for uncertainty emulation and characterization.

This minisymposium focuses on recent advances in uncertainty quantification algorithmic developments and applications based on data-driven and machine learning approaches in large-scale applications. Topics include data-driven surrogate construction, data assimilation, and physics-informed machine learning based on a limited number of data/observations and provide guidance for the system design, forecasting, etc.

14:00

Big Data Assimilation in Numerical Weather Prediction and Perspectives toward DA-UQ Collaboration

14:30

- CANCELED - Random regularization for adversarial robustness of neural networks

15:00

Trainability and Data-dependent Initialization of Over-parameterized ReLU networks

15:30

- CANCELED - Stochastic Training of Residual Networks: a Differential Equation Viewpoint

14:00

iCal
Michaël Baudin | EDF | France

Vladimir Cerisano Kovačević | Kobe Innovation Engineering and University of Florence | Italy

Sophie Ricci | CECI, CERFACS/CNRS-5318 | France

MS561: The OpenTURNS software for Uncertainty Quantification

Room:
MW 2250

Topic:
Uncertainty propagation

Form of presentation:
Mini-symposium

Duration:
120 Minutes

OpenTURNS is an open source library for uncertainty propagation by probabilistic methods. Developed by a partnership of five industrial companies (EDF, Airbus, Phimeca, IMACS and ONERA), it benefits from a strong practical feedback. Classical algorithms of UQ are available: central dispersion, probability of exceedance, sensitivity analysis, metamodels and stochastic processes. Developed in C++, OpenTURNS is also available as a Python module and has gained maturity thanks to more than 10 years of development. The goal of this minisymposium is to gather the OpenTURNS community and get an overview of the trends within the software, the associated research topics and its industrial uses.

14:00

Overview of OpenTURNS, its new features and its graphical user interface

14:30

Uncertainties in Civil Engineering: a FE-model calibration process for Cultural Heritage applications

15:00

Using OpenTURNS for surrogate modeling in the context of uncertainty reduction and data assimilation for 1D and 2D hydrodynamics

14:00

iCal
Qiqi Wang | Massachusetts Institute of Technology | United States

Trung Pham | University of Michigan | United States

Domenico Quagliarella | Centro Italiano Ricerche Aerospaziali (CIRA) | Italy

Fabio Nobile | Ecole Polytechnique Fédérale de Lausanne (EPFL) | Switzerland

MS512: UQ for complex fluid dynamics problems in realistic applications (Part II of II)

Chair(s)
Lorenzo Tamellini (Istituto di Matematica Applicata e Tecnologie Informatiche (CNR-IMATI))

Matteo Diez (Istituto di Ingegneria del Mare (CNR-INM))

Maria Vittoria Salvetti (University of Pisa)

Matteo Diez (Istituto di Ingegneria del Mare (CNR-INM))

Maria Vittoria Salvetti (University of Pisa)

Room:
MW 1701

Topic:
UQ for complex systems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

Uncertainty Quantification techniques are by now mature enough to address realistic, large scale problems of significant relevance. In this minisymposium, we focus in particular on complex fluid dynamics problems for engineering and environmental applications, that are fields in which computational science has traditionally played a major role. Several different kinds of UQ analyses naturally arise in these fields: forward UQ, optimization under uncertainty, inverse problem and data assimilation (e.g. for real-time control). In these scenarios, non-standard randomness might occur, and the complexity of the governing PDE equations further introduces significant and fascinating theoretical and computational challenges. In particular, polynomial-based UQ methods like Polynomial Chaos or Sparse grids collocation might not work well, in which case one has to resort to sampling methods, control variates, and more recently, machine learning techniques. Ad-hoc algorithms for high-performance computing are also relevant in this framework and welcome in this minisymposium.

14:00

Adaptive sampling for constructing complex aerodynamic response surfaces on the NASA High-Lift Common Research Model

14:30

Importance sampling and approximate control variates for complex physical simulation

15:00

Efficient Robust Design Optimization using Exploratory Data Analysis

15:30

- NEW - A multilevel stochastic gradient algorithm for PDE-constrained optimal control problems under uncertainty

14:00

iCal
Jakob Macke | Technical University of Munich | Germany

Casey Dowdle | Cold Regions Research and Engineering Laboratory | United States

Giulio Trigila | Baruch College of New York | United States

MS752: Advances in likelihood-free inference (Part II of II)

Chair(s)
Ricardo Baptista (Massachusetts Institute of Technology)

Giulio Trigila (Baruch College of New York)

Giulio Trigila (Baruch College of New York)

Room:
IAS 4.001

Topic:
Inverse problems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Many statistical models of interest in engineering, the sciences, and machine learning define a likelihood function that is computationally prohibitive to evaluate. This may be induced from the model only being known through a data generating process or the likelihood function involving a high-dimensional integral (e.g., from a marginalization procedure or the computation of a normalizing constant). In these cases, it is difficult to apply classical inference methods such as maximum likelihood estimation or likelihood-based Bayesian inference algorithms. To enable inference in these settings, several approaches have been developed in the statistics and machine learning community that avoid direct evaluation of the likelihood function (e.g., approximate Bayesian computation). Despite these success, efficiently solving such problems remains challenging, especially in high dimensions, or when only limited information or few samples are available. This mini-symposium will explore new algorithms and methodologies for performing likelihood-free inference in these complex models.

14:30

Sequential Neural Posterior Estimation for Likelihood-Free Inference

15:00

Accelerating inference with measure transport and generative networks

15:30

Characterization and simulation of conditional probability densities

14:00

iCal
Oindrila Kanjilal | Technical University of Munich | Germany

Jean-Marc Bourinet | SIGMA Clermont | France

Maliki Moustapha | ETH Zurich | Switzerland

Ziqi Wang | Guangzhou University | China

MS362: Theory and simulation of failure probabilities and rare events (Part II of II)

Chair(s)
Iason Papaioannou (Technical University of Munich)

Michael D. Shields (Johns Hopkins University)

Dimitrios Giovanis (Johns Hopkins University)

Elisabeth Ullmann (Technical University of Munich)

Michael D. Shields (Johns Hopkins University)

Dimitrios Giovanis (Johns Hopkins University)

Elisabeth Ullmann (Technical University of Munich)

Room:
Interims Lecture Hall 101

Topic:
Rare events and Risk

Form of presentation:
Mini-symposium

Duration:
120 Minutes

Organized in co-operation with

GAMM AG UQ

The evaluation of failure probabilities is a fundamental problem in reliability analysis and risk management of systems with uncertain inputs. We consider systems described by PDEs with random coefficients together with efficient approximation schemes. This includes stochastic finite elements, collocation, reduced basis, and advanced Monte Carlo methods. Efficient evaluation and updating of small failure probabilities and rare events remains a significant computational challenge. This mini-symposium brings together tools from applied probability, numerical analysis, and computational science and engineering. We showcase advances in analysis and computational treatment of rare events and failure probabilities, including variance reduction, advanced meta-models, and multilevel Monte Carlo.

14:00

Cross entropy-based importance sampling for first passage probability estimation of linear structures with parameter uncertainties

14:30

- CANCELED - Kernel-based adaptive models with tuned regularity parameters for rare-event probability estimation

15:00

Towards a global framework for reliability analysis based on active learning

15:30

- CANCELED - An active learning-based Gaussian process metamodelling strategy for estimating the probability distribution in forward UQ analysis

14:00

iCal
Geir Evensen | NORCE and Nansen Environmental and Remote Sensing Center | Norway

Alfredo Garbuno Inigo | California Institute of Technology | United States

Corinna Maier | University of Potsdam | Germany

Kody Law | University of Manchester | United Kingdom

MS772: Ensemble & Particle methods for inverse problems (Part II of II)

Room:
Interims Lecture Hall 102

Topic:
Inverse problems

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The understanding and incorporation of data within models has become a vital component of applied mathematics. A fundamental one can ask is given noisy measurements of data, how to recover some unknown quantity of interest. Some examples of these fields include in- verse problems which is primarily concerned with parameter estimation and data assimilation for state estimation. Both fields have seen a considerable amount of attention due to recent advancements in terms of both classical and statistical approaches. In particular, this mini-symposium will consider particle methods for solving inverse problems with the help optimization tools as well as particle methods aiming to represent the posterior distribution in a bayesian point of view for inverse problems.

The motivation behind this mini-symposium is to bring together experts from both schools. This would provide a complimentary field to the mini-symposium where connections between both areas are currently being developed.

14:00

Ensemble-subspace formulation of an iterative smoother for solving inverse problems

14:30

Optimize, Learn, Sample

15:00

Linking data assimilation with reinforcement learning for individualized dosing policies

15:30

Multilevel Monte Carlo methods for Bayesian inverse problems

14:00

iCal
Udo von Toussaint | Max Planck Institute for Plasmaphysics | Germany

James Oreluk | Sandia National Laboratories | United States

Sophie Harbisher | Newcastle University | United Kingdom

Andrew D. Davis | U.S. Army Corps of Engineers, Cold Regions Research and Engineering Laboratory | United States

MS892: Model-based Optimal Experimental Design (Part II of II)

Room:
Exzellenzzentrum 0003

Topic:
Design of experiments

Form of presentation:
Mini-symposium

Duration:
120 Minutes

The challenge of acquiring the most valuable data from experiments—for the purpose of inference, prediction, design, or control—has received substantial attention in statistics, applied mathematics, and engineering and science. This task can be formalized through the framework of optimal experimental design (OED). Models describing experimental conditions and processes, both physical and statistical, can be particularly useful for arriving at these optimal designs. However, model-based OED faces many challenges, such as formulational difficulties, choices of optimality criteria, computation of information metrics, handling nonlinear responses and non-Gaussian distributions, and dealing with expensive and dynamically evolving simulations. This minisymposium invites researchers of model-based OED, in the broad areas of computational and applications-oriented developments.

14:00

Optimal Bayesian Design of Thermal Desorption (TDS) and Temperature programmed Desorption (TPD) Experiments

14:30

Bayesian optimal experimental design for chemical rate constant measurement using mass spectrometry

15:00

High dimensional optimal design using stochastic gradient optimisation and Fisher information gain

15:30

Asymptotically exact optimal experimental design using local surrogate models to design aquifer monitoring strategies