MCQMC2014 — April 6 – 11, 2014 — KU Leuven, Belgium

Eleventh International Conference on
Monte Carlo and Quasi-Monte Carlo Methods in Scientific Computing

List of sessions

Click on the title of a talk to see the abstract. Click on [slides] to get at the slides. If the slides link is colored orange then they are password protected.

1. Tutorials
  1. Sun, Room A, 15:00-16:00
    ANOVA, global sensitivity, Sobol' indices and all that [slides]
    Art Owen (Stanford University)
  2. Sun, Room A, 16:00-17:00
    An introduction to multilevel Monte Carlo methods [slides]
    Mike Giles (University of Oxford)

2. Plenary talks
  1. Mon, Room A, 09:00-09:55
    Multilevel Monte Carlo for Lévy-driven SDEs [[slides]]
    Steffen Dereich (Westfälische Wilhelms-Universität Münster)
  2. Mon, Room A, 14:00-14:55
    Creating unbiased Monte Carlo schemes from biased ones: Theory and applications [slides]
    Peter Glynn (Stanford University)
  3. Tue, Room A, 09:00-09:55
    Walsh Figure of Merit (WAFOM) for digital nets: An easy measure for higher order convergent QMC [slides]
    Makoto Matsumoto (Hiroshima University)
  4. Wed, Room A, 09:00-09:55
    Some results on the complexity of numerical integration [slides]
    Erich Novak (Friedrich Schiller University Jena)
  5. Thu, Room A, 09:00-09:55
    Path space MCMC methods in computer graphics
    Wenzel Jakob (ETH Zürich)
  6. Thu, Room A, 14:00-14:55
    Adaptive strategies for Multilevel Monte Carlo
    Raul Tempone (King Abdullah University of Science and Technology)
  7. Fri, Room A, 09:00-09:55
    Fully automated Approximative Bayesian Computation [[slides]]
    Christian Robert (Université Paris-Dauphine)
  8. Fri, Room A, 14:00-14:55
    Vandermonde nets: A new family of digital nets [slides]
    Harald Niederreiter (Austrian Academy of Sciences)

3. Public lecture
  1. Tue, Maria Theresia College, 20:00-21:00
    Mathematical Imagery: A tale about Escher, the fourth dimension and chaos
    Jos Leys (Belgium)

4. Multilevel Monte Carlo for stochastic differential equations (Steffen Dereich, Mike Giles)
Keywords: Multi-Level Monte Carlo Methods (MLMC) Stochastic Differential Equations (SDEs) Numerical Analysis Novel applications Quasi-Monte Carlo (QMC) Methods

The development of multilevel methods has led to great reductions in the computational complexity of Monte Carlo simulation for stochastic (partial) differential equations. Central to it is the paradigm to work with a whole hierarchy of approximations instead of a single fixed one in numerical schemes.

Recent progress has been made in various directions:

  1. The variety of underlying processes has been significantly extended, including stochastic partial differential equations and backward stochastic differential equations.
  2. The theory has been extended to various cases where the standard theory does not apply, because of deficient regularity of the coefficients or payoff functionals.
  3. Adaptivity and Markov chain techniques have been combined with multilevel methods.

  1. Mon, Room A, 10:00-10:25
    From rough path estimates to multilevel Monte Carlo [slides]
    Christian Bayer (Weierstrass Institute)
  2. Mon, Room A, 10:25-10:50
    Multilevel Monte Carlo smoothing for approximation of densities [slides]
    Klaus Ritter (Technische Universität Kaiserslautern)
  3. Mon, Room A, 11:15-11:40
    Multilevel Monte Carlo for Lévy-driven SDEs: central limit theorems [slides]
    Sangmeng Li (Westfälische Wilhelms-Universität Münster)
  4. Mon, Room A, 11:40-12:05
    Customized tamed numerical schemes for SDEs and BSDEs [slides]
    Lukasz Szpruch (University of Edinburgh)
  5. Mon, Room A, 12:05-12:30
    Approximations and exponential moments for nonlinear stochastic differential equations [slides]
    Martin Hutzenthaler (University of Frankfurt)
  6. Mon, Room A, 15:00-15:25
    Multilevel quadrature of discontinuous payoffs in the generalized Heston model: Derivation [slides]
    Martin Altmayer (Universität Mannheim)
  7. Mon, Room A, 15:25-15:50
    Multilevel quadrature of discontinuous payoffs in the generalized Heston model: Error analysis [[slides]]
    Andreas Neuenkirch (Universität Mannheim)
  8. Mon, Room A, 16:15-16:40
    Adaptively balanced parallel MLMC solver for acoustic wave propagation with log-normal coefficients [slides]
    Jonas Sukys (ETH Zürich)
  9. Mon, Room A, 16:40-17:05
    Fault tolerant multilevel Monte Carlo [slides]
    Stefan Pauli (ETH Zürich)
  10. Mon, Room A, 17:05-17:30
    Multilevel simulation in a hybrid variance reduction method for derivative pricing
    Christoph Reisinger (University of Oxford)
  11. Tue, Room A, 10:00-10:25
    Multilevel Monte Carlo Methods for elliptic equations
    Andrea Barth (ETH Zürich)
  12. Tue, Room A, 10:25-10:50
    Multilevel Monte Carlo method with applications to stochastic partial differential equations [[slides]]
    Annika Lang (Chalmers University of Technology)
  13. Tue, Room A, 11:15-11:40
    Multilevel Monte Carlo methods for solving high-dimensional optimal stopping problems [slides]
    Denis Belomestny (Duisburg-Essen University)
  14. Tue, Room A, 11:40-12:05
    Variance reduction technique with sieve minimum variance approach for MLMC
    Tigran Nagapetyan (Fraunhofer ITWM)
  15. Tue, Room A, 12:05-12:30
    A multilevel approach to nested-simulation based risk measure estimation [slides]
    Yanchu Liu (The Chinese University of Hong Kong)
  16. Tue, Room A, 14:00-14:25
    Optimal selection of levels for multilevel Monte Carlo [slides]
    Mike Giles (University of Oxford)
  17. Tue, Room A, 14:25-14:50
    A continuation multilevel Monte Carlo algorithm [slides]
    Abdul-Lateef Haji-Ali (King Abdullah University of Science and Technology)
  18. Tue, Room A, 14:50-15:15
    Weak approximation of SDE by a strong error adaptive multilevel Monte Carlo method [slides]
    Haakon Hoel (King Abdullah University of Science and Technology)
  19. Tue, Room A, 15:15-15:40
    Multilevel Monte Carlo Chernoff tau-leap [slides]
    Alvaro Moraes (King Abdullah University of Science and Technology)
  20. Tue, Room A, 16:05-16:30
    Multilevel methods for mathematical biology
    Christopher Lester (University of Oxford)
  21. Tue, Room A, 16:30-16:55
    Multilevel Markov chain Monte Carlo algorithms for uncertainty quantification [slides]
    Aretha Teckentrup (Florida State University)
  22. Tue, Room A, 16:55-17:20
    A multilevel algorithm for backward stochastic differential equations [slides]
    Plamen Turkedjiev (Ecole Polytechnique)

5. Monte Carlo and Markov chain Monte Carlo for direct and inverse boundary value problems for PDE (Sylvain Maire, Martin Simon)
Keywords: Stochastic Processes Monte Carlo (MC) and Markov Chain Monte Carlo Methods (MCMC) Lagrangian methods Variance reduction methods Parallelization Applications

Probabilistic methods for solving boundary value problems for partial differential equations (PDEs) as well as the corresponding inverse problems have received increasing attention in the last 15 years, mainly due to the advent of multicore computing architectures. As a result, stochastic analysis and stochastic numerics have become important tools in both fields.

MC methods for direct problems (which may possibly depend on random parameters) are an interesting alternative to deterministic methods such as the finite element method, particularly when one needs to compute the solution at only a few points, or when moderate accuracy is sufficient. On top of that, their approximation error may be estimated on the fly. The corresponding inverse parameter identification problems are notoriously ill- posed and thus need to be regularized, e.g. by statistical inversion, where computing common estimators corresponds to MCMC sampling of a posterior probability density. However, due to high complexity, many practical applications require model reduction, careful error modeling and sampling schemes specifically tailored to reduce the computational cost. Recently novel MCMC sampling schemes that use a hierarchy of MC methods for the direct problem were developed.

This Special Session focuses on recent progress with regard to theoretical and numerical aspects in both, direct and inverse problems. All of the talks will also cover real life applications such as electrical impedance tomography, geophysical simulation of solute transport and geodynamic imaging.

  1. Mon, Room B, 10:00-10:25
    Constraining rock properties of the Earth through joint geodynamic inversion
    Tobias Baumann (University of Mainz)
  2. Mon, Room B, 10:25-10:50
    Improved Monte Carlo simulations of diffusions in stratified media
    Sylvain Maire (Université de Toulon)
  3. Mon, Room B, 11:15-11:40
    Electrical impedance tomography with probabilistic forward models
    Martin Simon (University of Mainz)
  4. Mon, Room B, 11:40-12:05
    PALMTREE: Parallel Add-on for Lagrangian Methods and Tracker Replay with Extreme Ease [slides]
    Lionel Lenotre (INRIA Rennes Bretagne -- Atlantique and Rennes 1)
  5. Mon, Room B, 12:05-12:30
    Variance reduction in probabilistic domain decomposition [[slides]]
    Francisco Bernal (Instituto Superior Tecnico)

6. Sequential Monte Carlo, rare events, and exact simulation (Jose Blanchet, Peter Glynn)
Keywords: Sequential Monte Carlo Island Particle Model Rare event simulation Large deviations Stochastic partial differential equations Perfect simulation Multilevel Monte Carlo

The talks of this session encompass topics ranging from sequential Monte Carlo for Bayesian inference, efficient importance sampling for rare event analysis, and topics at the interface of exact simulation and multilevel Monte Carlo.

  1. Mon, Room B, 15:00-15:25
    Unbiased simulation of distributions with explicitly known integral transforms
    Yiwei Wang (The Chinese University of Hong Kong)
  2. Mon, Room B, 15:25-15:50
    Sequential quasi-Monte Carlo [slides]
    Mathieu Gerber (University of Lausanne)
  3. Mon, Room B, 16:15-16:40
    The island particle filter: a step toward a massive particle filter [slides]
    Eric Moulines (Télécom ParisTech)
  4. Mon, Room B, 16:40-17:05
    Rare-event simulation for random elliptic differential equations
    Jingchen Liu (Columbia University)
  5. Mon, Room B, 17:05-17:30
    Exact sampling and multilevel Monte Carlo for multidimensional reflected Brownian motion [[slides]]
    Jose Blanchet (Columbia University)

7. Construction of QMC point sets (Josef Dick, Dirk Nuyens)
Keywords: Quasi-Monte Carlo (QMC) Methods Digital Nets, Lattice Rules, and Sequences Low-discrepancy Points and Sequences Component-by-Component Construction

Quasi-Monte Carlo rules are equal weight quadrature rules for approximating integrals over the unit cube. The set of quadrature points therefore completely determines the properties of the QMC rule, which highlights the importance of good construction of point sets. The order of convergence that can be obtained depends on the smoothness of the functions, typically the order of convergence is $O(N^{-1+\delta})$ but recently the focus is also on higher-order convergence $O(N^{-\alpha})$ with $\alpha > 1$. Another recent focus is on numerical integration of functions which have infinitely many coordinates. In this session various constructions of suitable point sets and sequences will be discussed.

  1. Mon, Room C, 10:00-10:25
    Refreshing {\sc MinT} [slides]
    Wolfgang Ch. Schmid (University of Salzburg)
  2. Mon, Room C, 10:25-10:50
    New implementations of global function field based constructions of qMC point sets [slides]
    Isabel Pirsic (Johannes Kepler University Linz)
  3. Mon, Room C, 11:15-11:40
    Korobov's algorithms for lattice rules [slides]
    Roel Matthysen (KU Leuven)
  4. Mon, Room C, 11:40-12:05
    The Hermite normal form for circulant and skew-circulant lattice rules [slides]
    Stephen Joe (The University of Waikato)
  5. Mon, Room C, 12:05-12:30
    Trigonometric interpolation on lattice grids [slides]
    Tor Sørevik (University of Bergen)
  6. Mon, Room C, 16:15-16:40
    The mean square quasi-Monte Carlo error for digitally shifted point sets part I [slides]
    Kosuke Suzuki (The University of Tokyo)
  7. Mon, Room C, 16:40-17:05
    The mean square quasi-Monte Carlo error for digitally shifted point sets part II [slides]
    Ryuichi Ohori (The University of Tokyo)
  8. Mon, Room C, 17:05-17:30
    Low-discrepancy quadrature in the triangle [slides]
    Kinjal Basu (Stanford University)
  9. Tue, Room C, 10:00-10:25
    Bounds on the Walsh coefficients by dyadic difference and an improved figure of merit for QMC [slides]
    Takehito Yoshiki (The University of Tokyo)
  10. Tue, Room C, 10:25-10:50
    Constructing good higher order digital nets in reduced construction costs [slides]
    Takashi Goda (The University of Tokyo)
  11. Tue, Room C, 11:15-11:40
    Low-WAFOM point sets with small $t$-values [slides]
    Shin Harase (Tokyo Institute of Technology)
  12. Tue, Room C, 11:40-12:05
    Propagation rules for $(u,m,\boldsymbol{e},s)$-nets and $(u,\boldsymbol{e},s)$-sequences [slides]
    Peter Kritzer (Johannes Kepler University Linz)
  13. Tue, Room C, 12:05-12:30
    Constructing quasi-random points with the golden ratio sequence and the Hilbert space filling curve [slides]
    Colas Schretter (Vrije Universiteit Brussel)

8. Advances in Monte Carlo methods for kinetic equations (Giacomo Dimarco, Giovanni Samaey)
Keywords: Particle Methods Stochastic Processes Monte Carlo (MC) Methods Applications of Monte Carlo (MC) and Quasi-Monte Carlo (QMC) Methods Variance Reduction Methods

Kinetic equations, such as the Boltzmann equation and approximations such as the Boltzmann BGK equation, are ubiquitous in the study of systems of interacting particles. They model the evolution of a probability distribution of particles in a position-velocity phase space, resulting in a 6-dimensional time-dependent partial differential equation that is usually solved via a stochastic particle discretization. This special session considers computational issues present in this context: (i) the presence of statistical noise and appropriate variance reduction techniques; and (ii) the presence of multiple time scales and associated time discretization issues.

  1. Mon, Room D, 10:00-10:25
    Binary interaction algorithms for the simulation of self-organized systems [slides]
    Giacomo Albi (University of Ferrara)
  2. Mon, Room D, 10:25-10:50
    Hybrid and Moment Guided methods for Boltzmann type equations [slides]
    Giacomo Dimarco (University of Toulouse III)
  3. Mon, Room D, 11:15-11:40
    Simulating individual-based models of bacterial chemotaxis with asymptotic variance reduction
    Giovanni Samaey (KU Leuven)
  4. Mon, Room D, 11:40-12:05
    Scalable and quasi-contractive Markov coupling of Maxwell collisions
    Mathias Rousset (INRIA Paris -- Rocquencourt)

9. Deterministic Markov chain Monte Carlo (Josef Dick, Makoto Matsumoto, Daniel Rudolf, Houying Zhu)
Keywords: Monte Carlo (MC) Methods Markov Chain Monte Carlo (MCMC) Quasi-Monte Carlo (QMC) Methods Completely uniformly distributed Random and Pseudo-Random Numbers Discrepancy Low-discrepancy Points and Sequences Variance Reduction Methods

This session is devoted to the derandomization of Markov chain Monte Carlo (MCMC) which leads to Markov chain quasi-Monte Carlo (MCQMC) methods. If certain assumptions are satisfied, Quasi-Monte Carlo (QMC) methods have a higher order of convergence compared to Monte-Carlo methods. The question is whether MCQMC methods can also have a better rate of convergence compared to MCMC. This area is significantly influenced by recent theoretical results on QMC methods, MCMC, the interplay between random and pseudo-random numbers, the construction of completely uniformly distributed sequences and Monte Carlo ideas. The aim of this session is to bring experts of different research areas together to discuss recent results and further computational/theoretical challenges.

  1. Tue, Room C, 14:00-14:25
    Convergence rates for the Array-RQMC method [slides]
    Pierre L'Ecuyer (University of Montreal and INRIA Rennes)
  2. Tue, Room C, 14:25-14:50
    Some periodic sequences for deterministic Markov chain Monte Carlo
    Takuji Nishimura (Yamagata University)
  3. Tue, Room C, 14:50-15:15
    Discrepancy bounds for a deterministic acceptance-rejection sampler [slides]
    Houying Zhu (The University of New South Wales)
  4. Tue, Room C, 15:15-15:40
    Do we need arbitrary precision random variate generators in Monte Carlo simulation?
    Josef Leydold (Vienna University of Economics and Business)
  5. Thu, Room B, 15:00-15:25
    Herded Gibbs sampling [slides]
    Luke Bornn (Harvard University)
  6. Thu, Room B, 15:25-15:50
    Discrepancy bounds of Markov chain quasi-Monte Carlo
    Daniel Rudolf (Friedrich Schiller University Jena)
  7. Thu, Room B, 15:50-16:15
    Non-asymptotic error bounds for sequential MCMC methods in multimodal settings
    Nikolaus Schweizer (Saarland University)

10. Approximation theory and numerics for SDEs (Stefan Geiss, Andreas Neuenkirch)
Keywords: Approximation Theory SDEs with non-standard coefficients Levy Processes Backward SDEs

This session is devoted to quantitative approximation problems for all kinds of stochastic differential equations and related approximation problems, where in particular methods from stochastic analysis and approximation theory are combined. Among the studied equations are SDEs with irregular coefficients, backward SDEs and SDEs with Lévy noise, while among the approximation problems are the computation of exit times and optimal mean square approximation.

One of the main goals of this session is the design of efficient approximation schemes and the construction of the corresponding numerical algorithms.

  1. Tue, Room D, 10:00-10:25
    Time discretization of FBSDE with polynomial growth drivers and reaction-diffusion PDEs
    Goncalo dos Reis (Technical University Berlin)
  2. Tue, Room D, 10:25-10:50
    The Euler scheme's discrete exit time approximation converges strongly at a rate 1/2
    Bruno Bouchard (ENSAE-ParisTech)
  3. Tue, Room D, 11:15-11:40
    On the Malliavin derivative of Lévy driven BSDEs
    Christel Geiss (University of Innsbruck)
  4. Tue, Room D, 11:40-12:05
    Numerical integration of piecewise smooth functions using noisy information [slides]
    Paweł Morkisz (AGH University of Science and Technology)
  5. Tue, Room D, 12:05-12:30
    Approximation of stochastic integrals under logarithmic smoothness
    Stefan Geiss (University of Innsbruck)
  6. Fri, Room B, 10:00-10:25
    Exact simulation for SDEs with discontinuous drift [slides]
    Pierre Etore (Grenoble University)
  7. Fri, Room B, 10:25-10:50
    Stochastic differential equations with discontinuous drift
    Gunther Leobacher (Johannes Kepler University Linz)
  8. Fri, Room B, 11:15-11:40
    Optimal approximation of Skorohod integrals and Skorohod SDEs [slides]
    Peter Parczewski (Universität Mannheim)
  9. Fri, Room B, 11:40-12:05
    Minimal asymptotic errors for approximation of SDEs with time-irregular coefficients
    Paweł Przybyłowicz (AGH University of Science and Technology)
  10. Fri, Room B, 12:05-12:30
    Approximations and regularities for nonlinear stochastic ordinary and partial differential equations
    Arnulf Jentzen (ETH Zürich)

11. Rare event simulation with applications to power systems (Daan Crommelin, Wander Wadman)
Keywords: Rare Event Simulation Techniques Importance Sampling Importance Splitting Cross-entropy Method Power Systems Reliability Analysis Applications of Monte Carlo (MC) and Quasi-Monte Carlo (QMC) Methods

Rare event simulation comprises a collection of methods and techniques to accelerate the computation of rare events by Monte Carlo simulation. The methodological development of rare event simulation techniques is an active research topic. At the same time, these techniques are applied in fields varying from queueing systems and communication networks to computational chemistry. A rather new field of application is that of power systems reliability. This session will be devoted to rare event simulation, with special attention to power system applications.

Many modern societies have grown accustomed to a very reliable electricity supply by power transmission grids. Although events of a grid failure like blackouts or power curtailments are expected to be rare, they may have major impacts. Therefore, grid operators want to assess grid reliability by estimating the extent to which these events will occur. Monte Carlo simulation is a popular and robust technique to estimate grid reliability indices. For example, to estimate the probability $p$ of a grid failure, sources of uncertainty (such as failures of grid connections, generation outages or weather dependent generation) are simulated and the resulting system state is determined. After repeating this for $n$ samples, the ratio of samples $\hat p_n$ that exhibit a grid failure is an unbiased estimator for $p$. However, as is well-known, the relative variance of this estimator is

\begin{eqnarray} \frac{\mathop{\rm Var} \hat p_n}{p^2} = \frac{(1 - p)p}{np^2} = \frac{1 - p}{np}, \nonumber \end{eqnarray}

which diverges as $p \to 0$. Hence for very small $p$ and a fixed accuracy a Crude Monte Carlo simulation will require a prohibitively large number of samples. Similar problems of computational inefficiency are encountered with the estimation of other grid reliability indices (e.g. expected duration of failures).

Rare event simulation techniques are developed to reduce this computational burden. Two main categories can be distinguished: importance sampling and (importance) splitting. Both techniques aim to make the occurrence of rare events more likely; the former by sampling from an alternative distribution and the latter by resampling (splitting) paths that lead closer to the rare event set. Both techniques yield an unbiased estimator by properly weighting the simulation outputs, and they may dramatically reduce the relative variance of the estimator for a fixed workload. In the last few years, both approaches have been introduced for analyzing power grid reliability.

  1. Tue, Room D, 14:00-14:25
    Network reliability simulation with extension to Marshall-Olkin copula-based dependent failures
    Bruno Tuffin (INRIA Rennes Bretagne -- Atlantique)
  2. Tue, Room D, 14:25-14:50
    Efficient simulation of cascading blackouts using splitting
    John Shortle (George Mason University)
  3. Tue, Room D, 14:50-15:15
    Using importance sampling to evaluate power system voltage stability [slides]
    Magnus Perninge (Lund University)
  4. Tue, Room D, 15:15-15:40
    Simulation of power grid reliability indices using a splitting technique [[slides]]
    Wander Wadman (CWI Amsterdam)

12. Hyperbolic cross and high-dimensional approximation (Vladimir Temlyakov, Tino Ullrich)
Keywords: Hyperbolic cross Discrepancy Multivariate approximation

Efficient approximation and integration of multivariate functions is a crucial task for the numerical treatment of many multi-parameter real-world problems. Typically, the computation time of the algorithms grows exponentially in the number of variables. Therefore one is interested in reasonabĺe model assumptions and algorithms. Hyperbolic cross approximation is a powerful method which is closely linked to function classes with a bounded mixed derivative. Those functions appear naturally in probability and discrepancy theory. They further serve as a suitable framework for the treatment of the electronic Schrödinger equation. Over the last 50 years the subject has developed into a beautiful and practically useful theory with a significant number of important unsolved problems. This special session addresses the progress in the field of multivariate approximation with a special focus on hyperbolic cross approximation.

  1. Tue, Room E, 11:15-11:40
    Optimal QMC integration in $d$-variate Besov spaces with higher mixed smoothness [slides]
    Tino Ullrich (University of Bonn)
  2. Tue, Room E, 11:40-12:05
    Sparse grid integration in reproducing kernel Hilbert spaces
    Jens Oettershagen (University of Bonn)
  3. Tue, Room E, 12:05-12:30
    $N$-Widths, $\varepsilon$-dimensions and high-dimensional hyperbolic cross approximations [slides]
    Dung Dinh (Vietnam National University)
  4. Tue, Room E, 14:00-14:25
    Hierarchical dimension-adaptive machine learning [[slides]]
    Bastian Bohn (University of Bonn)
  5. Tue, Room E, 14:25-14:50
    Sampling numbers of Sobolev embeddings on the $d$-dimensional torus
    Winfried Sickel (Friedrich Schiller University Jena)
  6. Tue, Room E, 14:50-15:15
    Optimal sampling of $d$-variate periodic functions with respect to modified hyperbolic crosses [[slides]]
    Glenn Byrenheid (University of Bonn)
  7. Tue, Room E, 15:15-15:40
    Approximation of non-periodic functions [slides]
    Gowri Suryanarayana (KU Leuven)
  8. Tue, Room E, 16:05-16:30
    Reconstructing multivariate trigonometric polynomials from samples along rank-1 lattices
    Lutz Kaemmerer (TU Chemnitz)
  9. Tue, Room E, 16:30-16:55
    Approximation of multivariate periodic functions based on rank-1 lattice sampling
    Toni Volkmer (TU Chemnitz)
  10. Tue, Room E, 16:55-17:20
    Greedy algorithms in numerical integration [slides]
    Vladimir Temlyakov (University of South Carolina)

13. Tractability of integration and approximation (Aicke Hinrichs, Mario Ullrich)
Keywords: complexity and tractability of multivariate problems (IBC) high-dimensional integration and approximation

Approximation and integration are prime examples for the study of the complexity of multivariate and infinitely variate continuous problems. This special session is devoted to talks covering a wide range of topics dealing with tractability issues of such problems.

  1. Wed, Room A, 10:00-10:25
    On the randomized complexity of Banach space valued integration problems
    Stefan Heinrich (Technische Universität Kaiserslautern)
  2. Wed, Room A, 10:25-10:50
    Tractability of multivariate integration in Hermite spaces of analytic functions [slides]
    Christian Irrgeher (Johannes Kepler University Linz)
  3. Wed, Room A, 11:15-11:40
    Complexity of Banach space valued and parametric stochastic Ito integration
    Thomas Daun (Technische Universität Kaiserslautern)
  4. Wed, Room A, 11:40-12:05
    Integration of permutation-invariant functions [[slides]]
    Markus Weimar (Philipps-University Marburg)
  5. Wed, Room A, 12:05-12:30
    Approximation of ridge functions
    Jan Vybiral (Technical University Berlin)
  6. Wed, Room A, 14:00-14:25
    Approximation of ridge functions: Tractability results [slides]
    Sebastian Mayer (University of Bonn)
  7. Wed, Room A, 14:25-14:50
    Multilevel higher-order QMC Galerkin discretization for parametric operator equations [slides]
    Frances Kuo (The University of New South Wales)
  8. Wed, Room A, 14:50-15:15
    Approximation and tractability of Sobolev embeddings [slides]
    Thomas Kühn (Universität Leipzig)
  9. Thu, Room A, 15:00-15:25
    Complexity of oscillatory integration for univariate Sobolev spaces [slides]
    Mario Ullrich (Friedrich Schiller University Jena)
  10. Thu, Room A, 15:25-15:50
    Uniform weak tractability of non-homogeneous multivariate problems
    Pawel Siedlecki (University of Warsaw)
  11. Thu, Room A, 15:50-16:15
    Bernstein widths and lower bounds for the Monte Carlo error
    Robert J. Kunsch (Friedrich Schiller University Jena)
  12. Thu, Room A, 16:15-16:40
    Complexity of solving nonlinear equations in the deterministic, randomized and quantum settings
    Maciej Goćwin (AGH University of Science and Technology)
  13. Thu, Room A, 17:05-17:30
    Optimal algorithms for doubly weighted approximation and integration of smooth functions [slides]
    Leszek Plaskota (University of Warsaw)
  14. Thu, Room A, 17:30-17:55
    $L_2$- and $S_{p,q}^rB$-discrepancy of (order $2$) digital nets [slides]
    Lev Markhasin (University of Stuttgart)
  15. Thu, Room A, 17:55-18:20
    Tractability of function approximation problems with general kernels
    Xuan Zhou (Illinois Institute of Technology)

14. Mathematical aspects of Monte Carlo methods for molecular dynamics (Tony Lelièvre, Mathias Rousset, Giovanni Samaey)
Keywords: Markov Chain Monte Carlo (MCMC) Rare-Event Simulation Applications of Monte Carlo (MC) and Quasi-Monte Carlo (QMC) Methods Stochastic Processes

This session deals with sampling problems for molecular dynamics. In particular, we consider the accuracy of sampling and methods that can accelerate sampling in the presence of rare events.

  1. Wed, Room B, 10:00-10:25
    Rare event estimation for SPDEs
    Charles-Edouard Bréhier (Ecoles des Ponts)
  2. Wed, Room B, 10:25-10:50
    Fluctuation analysis of adaptive multilevel splitting [slides]
    Arnaud Guyader (University of Rennes)
  3. Wed, Room B, 11:15-11:40
    Simulation of rare events using adaptive protocols [slides]
    Carsten Hartmann (Freie Universität Berlin)
  4. Wed, Room B, 11:40-12:05
    Mathematical analysis of accelerated dynamics [slides]
    Tony Lelièvre (Ecole des Ponts)
  5. Wed, Room B, 12:05-12:30
    Reducing the discretization sampling bias using Langevin dynamics splitting methods [slides]
    Charles Matthews (University of Edinburgh)
  6. Wed, Room B, 14:00-14:25
    Computation of sensitivities to some parameter for the invariant measure of a diffusion [slides]
    Raphaël Roux (Université Pierre et Marie Curie)
  7. Wed, Room B, 14:25-14:50
    Networks of rare events studied by multiple state replica exchange transition interface sampling
    David Swenson (Universiteit van Amsterdam)

15. Sensitivity analysis of model output (Stefano Tarantola)
Keywords: Sobol' indices Correlated input factors Time dependent input factors Large dimensionality models Random balance designs Traffic modelling

Mathematical modellers and regulatory agencies worldwide share the belief that sensitivity analysis is a key ingredient of the quality of a model-based study. According to the European Commission "sensitivity analysis can be used to explore how the impacts of the options you are analyzing would change in response to variations in key parameters and how they interact". The term interaction is also used in the guidelines for modellers of the US environmental Protection Agency: " [sensitivity analysis] methods should preferable be able to deal with a model regardless to assumptions about a model’s linearity and additivity, consider interaction effects among input uncertainties [\ldots], and evaluate the effect of an input while all the other inputs are allowed to vary as well". In spite of this call for the use of global tools for sensitivity analysis, plenty of cases are found in the literature where local sensitivity analysis is employed. In local analyses factors’ importance is investigated by derivative of the output with respect to that input; such derivatives are informative only at the base point where they are computed, but do not provide for an exploration of the rest of the space of the input factors unless some conditions are met in the mathematical formulation under analysis. In this session we treat new approaches to global sensitivity analysis and an application to traffic simulation modeling.

  1. Wed, Room C, 10:00-10:25
    Uncertainty management in traffic modelling [slides]
    Biagio Ciuffo (Joint Research Centre European Commission)
  2. Wed, Room C, 10:25-10:50
    Pick-freeze estimation of sensitivity indices for models with dependent causal processes inputs [slides]
    Mathilde Grandjacques (Grenoble University)
  3. Wed, Room C, 11:15-11:40
    Quasi-random balance designs for sensitivity analysis [slides]
    Stefano Tarantola (Joint Research Centre European Commission)
  4. Wed, Room C, 11:40-12:05
    Randomized pick-freeze for sparse estimation of Sobol indices in high dimension [slides]
    Alexandre Janon (Université Paris Sud)

16. QMC integration on the sphere, Riemannian manifolds and minimum energy problems (Johann Brauchart, Arno Kuijlaars)
Keywords: Quasi-Monte Carlo (QMC) Methods Low-discrepancy Points and Sequences QMC Designs Spherical Designs Minimum Energy Point Configurations Sphere Riemannian Manifold

Integrating a function over a manifold other than the unit cube usually involves a change of variable to map the domain back to the unit cube. Avoiding this step can be beneficial and motivates one main theme of this Special Session. A Quasi-Monte Carlo numerical integration method approximates the integral of a continuous function $f$ by means of the average of the function values at well-chosen nodes. One example on the sphere are QMC methods based on spherical $t$-designs which are exact for all spherical polynomials of degree at most $t$. Naturally, one can also use weighted averages of function values to approximate the integral. Reproducing kernel Hilbert space techniques provide a powerful and elegant method to directly compute the integration error and study its worst-case behavior. Discrete energy problems (e.g., optimal sum of mutual distances) emerge in a natural way when invoking this approach. But more general function space settings have also been studied.

This Special Session brings together experts and young researchers to discuss recent progress and the challenges of doing numerical integration using (weighted) QMC methods on the sphere and other manifolds, and investigates the connection to discrete minimum energy problems which can be used to characterize (QMC designs), find, and measure the quality of good node sets.

  1. Wed, Room C, 14:00-14:25
    QMC numerical integration on the sphere -- recent progress
    Johann Brauchart (The University of New South Wales)
  2. Wed, Room C, 14:25-14:50
    Discrepancy and numerical integration in Sobolev spaces on metric measure spaces
    Giacomo Gigante (University of Bergamo)
  3. Wed, Room C, 14:50-15:15
    Point sets of minimal energy
    Peter Grabner (Graz University of Technology)

17. Adaptive Markov chain Monte Carlo methods (Fabrizio Leisen)
Keywords: Adaptive Markov Chain Monte Carlo Methods Metropolis Algorithm Multiple-Try algorithm geometric ergodicity diminishing adaption condition containment condition Doeblin condition

Adaptive MCMC algorithms have recently become very popular since they often lead to significant speed-ups, even in high dimensional problems. For this reason, they became very popular in the last ten years and they have been used in many statistical applications. Despite their popularity, to set ergodic adaptive MCMC algorithms it's not easy since they violate the Markov property. Anyway, there are some theoretical results that can guarantie the convergence to the target distribution $\pi$. The aim of this session is to present some recent advances in adaptive MCMC with some active and pioneering researchers in the field.

  1. Wed, Room D, 10:00-10:25
    Assessing Monte Carlo errors in MCMC and adaptive MCMC
    Yves Atchade (University of Michigan)
  2. Wed, Room D, 10:25-10:50
    Enabling adaptation in pseudo-marginal Markov chain Monte Carlo
    Matti Vihola (University of Jyväskylä)
  3. Wed, Room D, 11:15-11:40
    Interaction and adaption via the multiple-try Metropolis
    Radu Craiu (University of Toronto)
  4. Wed, Room D, 11:40-12:05
    Sticky proposal densities for adaptive MCMC methods [slides]
    Luca Martino (University of Helsinki)

18. Markov chain Monte Carlo in high dimensions (Sergios Agapiou, Kasia Wolny)
Keywords: Markov chain Monte Carlo (MCMC) Large scale problems Sampling in high dimensions Gradient based methods

The problem of sampling high-dimensional distributions with possibly complex structure has received a lot of attention in recent years; in the Bayesian statistics context this has been driven on the one hand by the availability of more data and on the other hand by the desire for inference on more refined hence higher dimensional objects. It is well known that traditional MCMC algorithms deteriorate when the dimension of the state space is large and hence novel approaches are needed.

The aim of this special session is two-fold; to improve the understanding of the limitations of traditional MCMC methods in high dimensions, and most importantly to exchange ideas on new methods which are robust with respect to dimension. In particular we will discuss modifications of traditional MCMC algorithms which are motivated through function space intuition and are tailored to remain efficient in high-dimensional settings. The talks will also cover gradient based methods which exploit the local structure of the posterior distribution in order to efficiently sample distributions with heterogeneous scales.

  1. Wed, Room D, 12:05-12:30
    High dimensional analysis of the Gibbs sampler for hierarchical inverse problems [slides]
    Sergios Agapiou (University of Warwick)
  2. Wed, Room D, 14:00-14:25
    Function space analogues of common MCMC algorithms [slides]
    Simon Cotter (University of Manchester)
  3. Wed, Room D, 14:25-14:50
    Consistency and CLTs for stochastic gradient Langevin dynamics based on subsampled data [[slides]]
    Sebastian Vollmer (University of Oxford)
  4. Wed, Room D, 14:50-15:15
    Metropolis-adjusted Langevin algorithms with state-dependent scaling [[slides]]
    Kasia Wolny (University of Warwick)

19. Infinite-dimensional integration and quadrature of SDEs (Josef Dick, Michael Gnewuch)
Keywords: Infinite-dimensional integration Quasi-Monte Carlo (QMC) Methods Monte Carlo (MC) Methods Stochastic Differential Equations (SDEs) Stochastic Approximation Schemes Multilevel MC/QMC Algorithms Multivariate Decomposition Methods

Infinite-dimensional integration and quadrature of stochastic differential equations (SDEs) are challenging mathematical problems that are important for many practical applications. Both problems are linked. One may use, for instance, an infinite series representation of the driving stochastic process of the SDE at hand (as, e.g., the Lévy–Ciesielski or the Karhunen–Loéve representation of the Brownian motion). Computing the expectation of a functional depending on the solution of the SDE then turns into the problem of approximating an infinite-dimensional integral.

The session focusses on the complexity of quadrature of SDEs and infinite-dimensional integration and on efficient algorithms to tackle these problems. In the last few years exciting new methods, as, e.g., multilevel methods and multivariate decomposition methods, were introduced and lead to a significant speed up of algorithms and to (essentially) optimal complexity results for a number of problem settings. The talks of the session will present recent developments in this and related directions.

  1. Thu, Room A, 10:00-10:25
    A deterministic quadrature rule for marginals of SDEs based on weak order 2.0 Ito-Taylor steps
    Larisa Yaroslavtseva (University of Passau)
  2. Thu, Room A, 10:25-10:50
    Quadrature for self-similar distributions on $\mathbb{R}^d$
    Thomas Müller-Gronbach (University of Passau)
  3. Thu, Room A, 11:15-11:40
    Multivariate decomposition method (MDM) for infinite-dimensional integration [slides]
    Ian Sloan (The University of New South Wales)
  4. Thu, Room A, 11:40-12:05
    Integration w.r.t.\ the standard Gaussian measure on the sequence space [[slides]]
    Mario Hefter (Technische Universität Kaiserslautern)
  5. Thu, Room A, 12:05-12:30
    Deterministic and randomized polynomial lattice rules for integration on the sequence space [slides]
    Michael Gnewuch (Technische Universität Kaiserslautern)

20. Advanced light transport simulation (Alexander Keller)
Keywords: Monte Carlo (MC) Methods Quasi-Monte Carlo (QMC) Methods Markov Chain Monte Carlo (MCMC) Particle Methods Variance Reduction Methods Applications of Monte Carlo (MC) and Quasi-Monte Carlo (QMC) Methods

Synthesized images that cannot be distinguished from photographs are ubiquitous, especially in commercials, product design, and entertainment. While in early computer graphics approximating a look was sufficient, image synthesis nowadays requires the precise simulation of light transport and digital cameras based on physical principles. Such virtual measurements can be modeled as functionals of the solution of a second kind Fredholm integral equation.

Topics of the special session are the latest developments in advanced Monte Carlo and quasi-Monte Carlo light transport simulation algorithms.

  1. Thu, Room B, 10:00-10:25
    Consistent MC and MCMC integration of singular integrands using selective mollification
    Anton S. Kaplanyan (Karlsruhe Institute of Technology)
  2. Thu, Room B, 10:25-10:50
    Joint path sampling for rendering anisotropic participating media
    Iliyan Georgiev (Solid Angle Ltd.)
  3. Thu, Room B, 11:15-11:40
    At the boundary between light transport simulation and computational statistics
    Toshiya Hachisuka (Aarhus University)
  4. Thu, Room B, 11:40-12:05
    Combining path integral and particle density estimators in light transport simulation
    Jaroslav Křivánek (Charles University in Prague)
  5. Thu, Room B, 12:05-12:30
    Path space filtering for integro-approximation problems [[slides]]
    Alexander Keller (NVIDIA)

21. Discrepancy of QMC point sets (Peter Kritzer, Friedrich Pillichshammer)
Keywords: discrepancy uniform distribution modulo one special sequences diaphony

Finite and infinite point sets with excellent distribution properties are of high interest in the field of uniform distribution theory, and also as integration nodes in quasi-Monte Carlo algorithms. One way of assessing the quality of distribution of a given point set is to consider its discrepancy, which compares the relative number of points in subsets of the domain to the volume of these. Several forms of the discrepancy exist, most notably the extreme or the star discrepancy, and $L_p$-versions of these. In this special session, we would like to cover very recent results in discrepancy theory, ranging from theoretical aspects to computational methods. The special session tries to provide a mix of young researchers and leading experts in discrepancy theory, giving them the opportunity to present their current work and to discuss recent developments in this exciting research area.

  1. Thu, Room B, 17:05-17:30
    The discrepancy of random points
    Benjamin Doerr (Ecole Polytechnique)
  2. Thu, Room B, 17:30-17:55
    A survey and some improvements on low-discrepancy sequences [slides]
    Henri Faure (Aix-Marseille Université)
  3. Thu, Room B, 17:55-18:20
    Some new results on the discrepancy of point sequences [slides]
    Gerhard Larcher (Johannes Kepler University Linz)
  4. Fri, Room C, 10:00-10:25
    $LS$-sequences of points [slides]
    Ingrid Carbone (University of Calabria)
  5. Fri, Room C, 10:25-10:50
    The $L^1$ dichotomy for the discrepancy function estimates [slides]
    Dmitriy Bilyk (University of Minnesota)
  6. Fri, Room C, 11:15-11:40
    QMC and low discrepancy points in function spaces [slides]
    Aicke Hinrichs (University of Rostock)
  7. Fri, Room C, 11:40-12:05
    Explicit constructions of QMC rules for unweighted function spaces [slides]
    Josef Dick (The University of New South Wales)
  8. Fri, Room C, 12:05-12:30
    Explicit constructions of point sets and sequences with low $L_2$ discrepancy [slides]
    Friedrich Pillichshammer (Johannes Kepler University Linz)

22. Talks in honour of Ilya M. Sobol' (Sergei Kucherenko, Art Owen)
Keywords: Sobol' indices Global sensitivity analysis Sobol'-Hoeffding Anova decomposition Quasi-Monte Carlo (QMC) Methods

This is a special session to honour Ilya M. Sobol' in appreciation of his profound contributions to our field. Two of his many contributions are especially influential. He was a pioneer in digital quasi-Monte Carlo methods, developing the LP-tau sequences which yield his digital (t,s)-sequences in base 2. His more recent work in global sensitivity analysis centers on variance and derivative based measures. The former are now known as Sobol' indices. Those indices are a core technology in uncertainty quantification, an area whose importance increases along with the ever greater usage of computer models in science and industry. A list of Sobol's publications, up to 2006, appears in a note by M. K. Kerimov in Computational Mathematics and Mathematical Physics, volume 47, number 2, pages 1065-1072, 2007. This special session brings together eight researchers all deeply influenced by Sobol', presenting some of their current work related to Sobol's contributions.

  1. Thu, Room C, 10:00-10:25
    Higher order Sobol' indices [slides]
    Art Owen (Stanford University)
  2. Thu, Room C, 10:25-10:50
    On ANOVA decompositions of positive definite kernels and Gaussian random field paths [slides]
    David Ginsbourger (University of Bern)
  3. Thu, Room C, 11:15-11:40
    Derivative-based global sensitivity measures and their link with Sobol’ sensitivity indices [slides]
    Sergei Kucherenko (Imperial College London)
  4. Thu, Room C, 11:40-12:05
    Recent inference approaches for Sobol' sensitivity indices
    Clémentine Prieur (Grenoble University)
  5. Thu, Room C, 12:05-12:30
    Sensitivity indices beyond sensitivity analysis [slides]
    Andrea Saltelli (Joint Research Centre European Commission)
  6. Thu, Room C, 17:05-17:30
    The impact of effective dimension and discontinuity on the accuracy of quasi-Monte Carlo methods [[slides]]
    Xiaoqun Wang (Tsinghua University)
  7. Thu, Room C, 17:30-17:55
    Simulation of multivariate Poisson processes [[slides]]
    Alexander Kreinin (IBM)
  8. Thu, Room C, 17:55-18:20
    Reliable error estimation for cubature using Sobol' sequences [slides]
    Fred J. Hickernell (Illinois Institute of Technology)

23. Making MC faster: New advances in MC and MCMC (Kody Law, Raul Tempone)
Keywords: Monte Carlo (MC) Methods Markov Chain Monte Carlo (MCMC)

Sampling from high-dimensional distributions can be prohibitively expensive, particularly if the underlying algorithm degenerates in the limit as dimension tends to infinity. Even if the algorithm is defined on function space, issues such as a broad range of scales and multiple modes limit the efficiency. Furthermore, constraints on the distribution or observables can introduce further difficulties. This special session aims to bring together experts in the field to share their most recent ideas about how to overcome these difficulties to make Monte Carlo sampling algorithms faster.

  1. Thu, Room D, 15:00-15:25
    On non-negative unbiased estimators [slides]
    Pierre Jacob (University of Oxford)
  2. Thu, Room D, 15:25-15:50
    A function space HMC algorithm with second order Langevin diffusion limit
    Michela Ottobre (Imperial College London)
  3. Thu, Room D, 15:50-16:15
    Proposals which speed up function-space MCMC samplers
    Kody Law (King Abdullah University of Science and Technology)
  4. Thu, Room D, 17:05-17:30
    Hybrid Chernoff tau-leap
    Pedro Vilanova (King Abdullah University of Science and Technology)
  5. Thu, Room D, 17:30-17:55
    Parallel MCMC [slides]
    Scott Schmidler (Duke University)

24. MC and QMC in finance (Giray Okten)
Keywords: Computational finance Monte Carlo and quasi-Monte Carlo methods

Monte Carlo and quasi-Monte Carlo methods are popular numerical tools in computational finance. In the past we have seen several research problems in Monte Carlo motivated by examples from financial mathematics. The talks in this special session will present recent research and applications in financial mathematics. In particular, the talks will cover new stochastic models as well as efficient implementations of existing ones, an investigation of model robustness in the context of weather derivatives modeling, GPU computing for financial applications, and an acceptance rejection method for low-discrepancy sequences and its potential impact on financial simulation.

  1. Thu, Room E, 15:00-15:25
    The acceptance-rejection method for low-discrepancy sequences and GPU computing [slides]
    Giray Okten (Florida State University)
  2. Thu, Room E, 15:25-15:50
    Efficient implementation of the variance gamma model using a QMC acceptance-rejection algorithm [slides]
    Nguyet Nguyen (Florida State University)
  3. Thu, Room E, 15:50-16:15
    Global sensitivity analysis in weather derivatives pricing [slides]
    Ahmet Goncu (Xian Jiaotong Liverpool University)
  4. Thu, Room E, 16:15-16:40
    Option pricing with CAM stochastic volatility model using Monte Carlo [[slides]]
    Wanwan Huang (Roosevelt University)
  5. Thu, Room E, 17:05-17:30
    Hybrid-Monte Carlo methods for credit risk management
    Lucia Del Chicca (Johannes Kepler University Linz)
  6. Thu, Room E, 17:30-17:55
    Optimal portfolio executions: a Monte Carlo approach [slides]
    Nico Achtsis (KU Leuven)
  7. Thu, Room E, 17:55-18:20
    Efficient implementation of Markov chain parameter estimation algorithms on GPU [slides]
    Emanouil Atanassov (Bulgarian Academy of Sciences)
  8. Fri, Room D, 10:00-10:25
    Efficient credit risk simulations in OpenCL [slides]
    Halis Sak (Yeditepe University)
  9. Fri, Room D, 10:25-10:50
    Determining market composition via principal component analysis of implied volatility surfaces
    Charles Joseph (Case Western Reserve University)
  10. Fri, Room D, 11:15-11:40
    Importance sampling for jump processes and applications to finance [slides]
    Jérôme Lelong (Grenoble University)
  11. Fri, Room D, 11:40-12:05
    Variance reduction procedures for financial risk estimation [slides]
    Mykhailo Pupashenko (Technische Universität Kaiserslautern)
  12. Fri, Room D, 12:05-12:30
    Monte Carlo methods for rare-event estimation with applications to insurance and risk management [slides]
    Rodrigo Targino (University College London)

25. Algorithmic aspects of integration (Alan Genz, Dirk Nuyens)
Keywords: Numerical integration Algorithms

This session looks at the practical and theoretical aspects of numerical integration using Monte Carlo and quasi-Monte Carlo algorithms.

  1. Fri, Room A, 10:00-10:25
    Numerical computation of multivariate normal probabilities using bivariate conditioning [slides]
    Alan Genz (Washington State University)
  2. Fri, Room A, 10:25-10:50
    Enhancing quasi-Monte Carlo methods via delta dimension and directional control technique [slides]
    Kai Liu (University of Waterloo)
  3. Fri, Room A, 11:15-11:40
    Error estimation for multidimensional integration based on rank-1 lattices [slides]
    Lluís Antoni Jiménez Rugama (Illinois Institute of Technology)
  4. Fri, Room A, 11:40-12:05
    Qint: Quasi-Monte Carlo integration algorithm with a posteriori error analysis [slides]
    Anton Antonov (Saint Petersburg State University)
  5. Fri, Room A, 12:05-12:30
    A guaranteed automatic integration library [slides]
    Lan Jiang (Illinois Institute of Technology)

26. Contributed talks
  1. Mon, Room D, 15:00-15:25
    Adaptive importance sampling using mixtures
    Tim Brereton (Universität Ulm)
  2. Mon, Room D, 15:25-15:50
    Particle flow for nonlinear filters, Bayesian decisions and transport [slides]
    Fred Daum (Raytheon)
  3. Mon, Room D, 16:15-16:40
    Monte-Carlo based multiple testing with bounded risk
    Georg Hahn (Imperial College London)
  4. Mon, Room D, 16:40-17:05
    A new rejection sampling method for truncated multivariate Gaussian random variables [slides]
    Hassan Maatouk (Ecole des Mines de St-Etienne)
  5. Mon, Room D, 17:05-17:30
    Strong law of large numbers and simulation of random LU-fuzzy numbers
    Tomas Tichy (VSB-TU Ostrava)
  6. Tue, Room B, 10:00-10:25
    Particle filtering approach in inference of biochemical pathways
    Vilda Purutçuoğlu (Middle East Technical University)
  7. Tue, Room B, 10:25-10:50
    The Wigner-Boltzmann Monte Carlo method and its applications to quantum computing [slides]
    Jean Michel Sellier (Bulgarian Academy of Sciences)
  8. Tue, Room B, 11:15-11:40
    Implementing hybrid PDE solvers [slides]
    Manolis Vavalis (Univeristy of Thessaly)
  9. Tue, Room B, 11:40-12:05
    A lattice Monte Carlo study of the aggregation of membrane protein TatA
    Yuanwei Xu (University of Warwick)
  10. Tue, Room B, 14:00-14:25
    Randomized QMC for weighted uniform sampling (WUS) with applications in quantum mechanics [slides]
    Hernan Leovey (Humboldt-University Berlin)
  11. Tue, Room B, 14:25-14:50
    A computational study of filter-based optimization algorithm by randomized quasi-Monte Carlo method [slides]
    Hozumi Morohosi (National Graduate Institute for Policy Studies)
  12. Tue, Room B, 14:50-15:15
    Sampling properties of Sudoku-based space-filling designs [slides]
    Duy Nguyen (University of Wisconsin-Madison)
  13. Tue, Room B, 15:15-15:40
    QMC methods are successful in mixed-integer stochastic optimization [slides]
    Werner Roemisch (Humboldt-University Berlin)
  14. Tue, Room B, 16:05-16:30
    Quasi-Monte-Carlo methods for the linear and nonlinear stochastic Poisson-Boltzmann equations [slides]
    Amirreza Khodadadian (TU Vienna)
  15. Tue, Room B, 16:30-16:55
    MC and QMC approaches for the numerical stochastic homogenization of elliptic PDEs
    Gerhard Tulzer (TU Vienna)
  16. Tue, Room B, 16:55-17:20
    An approximation of the weight distribution of the $n$-th bits of pseudorandom number generators [slides]
    Hiroshi Haramoto (Ehime University)
  17. Tue, Room C, 16:05-16:30
    MCMC for Bayesian analysis of accelerated life testing data of series systems [slides]
    Chiranjit Mukhopadhyay (Indian Institute of Science)
  18. Tue, Room C, 16:30-16:55
    Metropolis-type algorithms for hidden continuous time Markov processes [slides]
    Wojciech Niemiro (University of Warsaw)
  19. Tue, Room C, 16:55-17:20
    MCMC of stochastic matrices reversible with respect to a fixed stationary vector
    Benjamin Trendelkamp-Schroer (Freie Universität Berlin)
  20. Tue, Room D, 16:05-16:30
    Statistical error analysis for coagulation--advection simulations [slides]
    Robert Patterson (Weierstrass Institute)
  21. Tue, Room D, 16:30-16:55
    Bayesian conditional density filtering for big data
    Shaan Qamar (Duke University)
  22. Tue, Room D, 16:55-17:20
    Time series resampling for causality testing [slides]
    Angeliki Papana (University of Macedonia)
  23. Thu, Room C, 15:00-15:25
    Computing the Feynman-Kac formula efficiently with multilevel Monte Carlo [slides]
    Robert Gantner (ETH Zürich)
  24. Thu, Room C, 15:25-15:50
    Mixed precision multilevel Monte Carlo algorithms on reconfigurable hardware
    Steffen Omland (Technische Universität Kaiserslautern)
  25. Thu, Room C, 15:50-16:15
    High order weak extrapolation extension of the multilevel Monte Carlo method [slides]
    Anton Kostiuk (Technische Universität Kaiserslautern)
  26. Thu, Room C, 16:15-16:40
    Plasma physics applications and variance reduction for multilevel Monte Carlo [slides]
    Lee Ricketson (Univerisity of California, Los Angeles)
  27. Thu, Room D, 10:00-10:25
    Uncertainty propagation in the burnup Monte Carlo ALEPH code [slides]
    Alexey Stankovskiy (SCK-CEN)
  28. Thu, Room D, 10:25-10:50
    A new walk on equations Monte Carlo method for linear algebraic problems [slides]
    Ivan Dimov (Bulgarian Academy of Sciences)
  29. Thu, Room D, 11:15-11:40
    Continuous and discrete semimartingales decomposition and representation
    Antonio Dalessandro (University College London)
  30. Thu, Room D, 11:40-12:05
    Simulation of BSDEs by Wiener chaos expansion [slides]
    Céline Labart (Université de Savoie)
  31. Thu, Room D, 12:05-12:30
    Solutions to SDEs with discontinuous drift and singular diffusion
    Michaela Szölgyenyi (Johannes Kepler University Linz)

Revision: 226 Date: 2014-05-27 14:59:40 +0200 (Tue, 27 May 2014) Author: dirkn