HGS MathComp Curriculum & Events

The computerized revolution poses challenges and creates new opportunities in prehistorical research of the Southern Levant. The Computerized Archaeology Laboratory at the Hebrew University integrates techniques and ideas from computer science (e.g., computer graphics, machine learning) in the archaeological research methodologies. We operate optical scanners which provide three dimensional (3D) digital models that are analyzed with computer programs developed in our laboratory. We address research issues and needs which could not be addressed without the availability of digital 3D models. Beyond ‘capturing’ and visualizing data, I will focus on the process of analysis and provide novel ways of interpretation. Several examples will be presented as the clustering between prehistoric lithic assemblages while verifying or negating the traditional classification. This will provide new insights to the path of Homo Erectus out of Africa c.a. 1.4 million years ago and the Neanderthal occupation in the Southern Levant c.a. 100,000 years ago.

Data assimilation is the science of combining observations of a system with a numerical model of that system. It is used to initialise a model for prediction, like in numerical weather prediction, to provide a good estimate for the long evolution of a system, for instance in weather and climate reanalysis, and for model improvement via parameter estimation and sometimes even parameterisation estimation. It is used in all geosciences, but also in traffic control, neurosciences, agriculture, the process industry etc etc.

This course will provide a compact overview of the data-assimilation methods used, with emphasis on the geosciences. We will discuss traditional methods like 3DVar and 4DVar, Kalman Filters and their ensemble variants, but also new developments like hybrid methods, and fully nonlinear methods like particle filters. The lectures are accompanied by computer practicals in which the participants will get hands-on experience with the functioning of all these methods, and their pro’s and con’s. Use will be made of the EMPIRE data-assimilation system, which contains a whole suite of ensemble-based data-assimilation methods, and a whole suite of models used in the geosciences such as atmosphere models, ocean models, land-surface models, space weather models etc. Unique is the fast and efficient connection of any complex model with the EMPIRE data-assimilation system via MPI statements.

The outcome of the course will be a thorough understanding of the basics of present-day and potential future data-assimilation methods, their pro’s and con’s, and some experience with using them on systems of up to intermediate complexity.

Please register
here

Das Zusammenwirken von Physiologie, Orthopädie, Robotik, Regelungstechnik und Materialwissenschaft hat in den letzten Jahren einen rasanten Entwicklungsschub im Bereich der Prothetik und Orthetik ausgelöst. Das Seminar ermöglicht einen Blick auf die aktuelle Forschung im technischen und medizinischen Bereich und zeigt die Möglichkeiten der Anwendbarkeit in der heutigen Medizin. Geschichte, Ethik und Theologie erweitern den Reflektionsraum.
Ziel der Prothetik ist es, Patienten als Ersatz von fehlenden Körperteilen eine individuelle multifunktionale prothetische Versorgung zu bieten, die es ihnen ermöglicht, alltägliche Bewegungen natürlich und ungehindert auszuführen. Spezielle Prothesen, z.B. für bestimmte Sportarten, erlauben den Athleten inzwischen ein derartig hohes Leistungsniveau, dass ihnen teilweise ein Vorteil gegenüber Athleten ohne Prothesen unterstellt wird. Orthesen und Exoskelette hingegen ersetzen keine Gliedmaßen, sondern dienen existierenden Körperteilen Stütze oder Bewegungsantrieb. Ihr Einsatzbereich reicht von der lokalen Bewegungsunterstützung einzelner Gelenke bis zum Enhancement der Kräfte gesunder Menschen und zur Ganggenerierung bei Querschnittsgelähmten.
Vor 100 Jahren, im Ersten Weltkrieg, füllten Hunderte von Prothesen-Modellen die Ausstellungssäle populärer Kriegsausstellungen. Aus historischer Sicht werden technische, medizinische und soziale Entwicklungen und ihre Auswirkungen auf Patienten analysiert. In der Philosophie bestimmt die Enhancement-Debatte die Diskussion über Chancen und Risiken der Robotik. Aus ethischer Sicht wäre zu fragen, inwiefern Prothetik und Orthetik auch die Person als Ganzes betreffen und als Eingriff in die Persönlichkeit verstanden werden müssen. Diese Frage stellt sich insbesondere, wenn man Körper und Persönlichkeit nicht dualistisch trennt, sondern das menschliche Leben als prinzipiell verkörpertes versteht, so dass alle körperlichen Prozesse stets in einem Wirkungszusammenhang mit psychischen Prozessen stehen.

Die Anmeldung erfolgt über das Marsilius-Kolleg.

Schedule:

November 26, 2015, 13:00-17:00 - Theoretical foundations
November 27, 2015, 09:30-12:30 - Applications in image processing

Registration:

Please register at
http://hgs.iwr.uni-heidelberg.de/Portfolio_HGS/FGEN/form/fgen_form.php?id_form=31

Abstract:

Sparse stochastic processes are continuous-domain processes that admit a parsimonious representation in some matched wavelet-like basis. Such models are relevant for image compression, compressed sensing, and, more generally, for the derivation of statistical algorithms for solving ill-posed inverse problems.

The course will introduce the participants to the extended family of sparse processes that are specified by a generic (non-Gaussian) innovation model or, equivalently, as solutions of linear stochastic differential equations driven by white Lévy noise. We shall provide a complete functional characterization of these processes and highlight some of their properties. The two leading threads that underlie the exposition are:

1) the statistical property of infinite divisibility, which induces two distinct types of behavior Gaussian vs. sparse at the exclusion of any other;

2) the structural link between linear stochastic processes and splines.

The formalism lends itself to the derivation of the transform-domain statistics of these processes and to the identification of optimal (ICA-like) representations. We also show that these models are applicable to the derivation of statistical algorithms for solving ill-posed inverse problems, including compressed sensing. The proposed formulation leads to a reinterpretation of popular sparsity-promoting processing schemes such as total-variation denoising, LASSO, and wavelet shrinkage as MAP estimators for specific types of sparse processes, but it also suggests alternative Bayesian recovery procedures that minimize the estimation error.

The lecture notes for the course are available on the web at www.sparseprocesses.org.

About the instructor:

Michael Unser is Professor and Director of EPFL’s Biomedical Imaging Group, Lausanne, Switzerland. His main research area is biomedical image processing. He has a strong interest in sampling theories, multiresolution algorithms, wavelets, the use of splines for image processing, and, more recently, stochastic processes. He has published about 250 journal papers on those topics. He is the leading author of “An introduction to sparse stochastic processes“, Cambridge University Press, 2014.

From 1985 to 1997, he was with the Biomedical Engineering and Instrumentation Program, National Institutes of Health, Bethesda USA, conducting research on bioimaging and heading the Image Processing Group.

Dr. Unser is a fellow of the IEEE (1999), an EURASIP fellow (2009), and a member of the Swiss Academy of Engineering Sciences. He is the recipient of several international prizes including three IEEE-SPS Best Paper Awards and two Technical Achievement Awards from the IEEE (2008 SPS and EMBS 2010).

Schedule

Wednesday, December 16

14:00-15:30 Introduction to iterative solvers for large and sparse linear systems

15:30- 16:00 Coffee break
16:00-17:30 Preconditioning techniques

Thursday, December 17

9:00-10:30 Saddle-point systems: applications and spectral properties
10:30-11:00 Coffee break
11:00-12:30 Solution methods for saddle-point systems
12:30-14:00 Lunch break
14:00-16:30 Lab work

Friday, December 18

9:00-10:30 Eigenvalue solvers, part I: solvers based on decompositions
10:30-11:00 Coffee break
11:00-12:30 Eigenvalue solvers, part II: solvers based on matrix-vector products


The course concerns the numerical solution of two fundamental problems in numerical linear algebra: eigenvalue problems and linear systems. The first part describes algorithms for solving eigenvalue problems: QR iterations, methods for the symmetric tridiagonal eigenproblem, Lanczos and Arnoldi, Jacobi-Davidson, and other methods. In the second part we will discuss iterative Krylov subspace methods for solving large sparse linear systems, such as conjugate gradients, MINRES, and GMRES. We then turn our attention to block-structured linear systems arising from problems with constraints. Spectral properties of these matrices and the performance of various iterative solvers will be discussed, and we will cover in detail various preconditioning methodologies based on effective approximations of Schur complements.

Please register here

Many problems in machine learning that involve discrete structures or subset selection may be phrased in the language of submodular set functions. The property of submodularity, also referred to as a _discrete analog of convexity_, expresses the notion of diminishing marginal returns, and captures combinatorial versions of rank and dependence. Submodular functions occur in a variety of areas including graph theory, information theory, combinatorial optimization, stochastic processes and game theory. In machine learning, they emerge in different forms as the potential functions of graphical models, as the utility functions in active learning and sensing, in models of diversity, in structured sparse estimation or network inference.
The lectures will give an introduction to the theory of submodular functions, their applications in machine learning and related optimization problems.

Part I of the lectures will introduce the concept of submodularity along with several examples, as well as associated polyhedra and relations to convexity. Part II will address the ideas underlying algorithms for minimizing and maximizing submodular functions. For example, those algorithms exploit ties to both convexity and concavity.

Please register here

Krylov subspace methods are widely used for solving linear algebraic systems, approximating eigenvalues, and model reduction in science and engineering. This course will give a short overview of their mathematical foundations.

First, Krylov subspace methods use linear projections onto a sequence of finite dimensional subspaces. Since these subspaces are formed using repeated applications of the given matrix (operator) to a given initial vector, the methods are highly nonlinear. In fact, the nonlinearity allows for adaptation of the computation to the problem, which often results in an expected acceleration of convergence. If no substantial acceleration occurs and the results of the computation can be described using linear contractions, then one should ask whether the chosen Krylov subspace method is efficient for solving the given problem. The strong nonlinearity poses significant challenges for the analysis of the methods as well
as the understanding of effects of finite precision arithmetic in practical computations.

Different mathematical viewpoints and the resulting tools are used for analysis of Krylov subspace methods.The methods are closely linked with matching moments model reduction and quadrature. Spectral properties (eigenvalues and eigenvectors) of the matrix, together with the projections of the initial vector on the individual invariant subspaces, describe the convergence behavior, provided that the individual spectral invariant subspaces are mutually orthogonal. The general case is intriguing and still
far from being fully understood. Numerical stability of Krylov subspace methods is closely linked with the loss of orthogonality of the computed generating vectors, which is governed by different mechanisms for methods using short and long recurrences.

Finally, it is beneficial to consider formulations of Krylov subspace methods in infinite dimensional Hilbert spaces. That naturally leads to the concept of operator preconditioning. Preconditioning is a crucial part of computations using Krylov subspace methods; in most cases it is based on algebraic considerations and the discretized system of equations. Therefore it could be of interest to make a link between the operator and algebraic preconditioning.

The course will end with discussion of open questions.

Material used in this course:

J. Malek and Z. Strakos, Preconditioning and the Conjugate Gradient Method in the Context of Solving PDEs. SIAM Spotlight Series, SIAM (2015)

J. Liesen and Z. Strakos, Krylov Subspace Methods, Principles and Analysis. Oxford University Press (2013)

and the papers

G. Meurant and Z. Strakos, The Lanczos and conjugate gradient algorithms in finite precision arithmetic, Acta Numerica 15, 471-542 (2006)

T. Gergelits and Z. Strakos., Composite convergence bounds based on Chebyshev polynomials and finite precision conjugate gradient computations, Numer. Alg. 65, 759-782 (2014)

J. Papez, J. Liesen and Z. Strakos, Distribution of the discretization and algebraic error in numerical solution of partial differential equations, Linear Alg. Appl. 449, 89-114 (2014)

Z. Strakos and P. Tichy, On efficient numerical approximation of the bilinear form $c^* A^{-1} b$ , SIAM J. Sci. Comput. 33, 565-587 (2011)

J. Hench and Z. Strakos, The RCWA method - a case study with open questions and perspectives of algebraic computations , ETNA 31, 331-357 (2009)

Please register
here

The Distributed and Unified Numerics Environment (DUNE) is a software framework for the numerical solution of partial differential equations with grid-based methods. Using generic programming techniques it strives for both: high flexibility (efficiency of the programmer) and high performance (efficiency of the program). DUNE provides, among other things, a large variety of local mesh refinement techniques, a scalable parallel programming model, an ample collection of finite element methods and efficient linear solvers.

DUNE-PDELab is a powerful tool for implementing discretisations of partial-differential equations. It helps to substantially reduce the time to implement discretizations and solvers for (systems of) PDEs based on DUNE. It is not only suitable for rapid prototyping but also for building highly performant simulation software and is used by a variety of projects already.

This one week course will provide an introduction to the most important DUNE modules and especially to DUNE-PDELab. At the end the attendees will have a solid knowledge of the simulation workflow from mesh generation and implementation of finite element and finite volume methods to visualization of the results. Topics covered are the solution of stationary and time-dependent problems, as well as local adaptivity, the use of parallel computers and the solution of non-linear PDE_s and systems of PDE_s.

Wir haben die Daten - nutzen wir sie!

Nach Public Domain und Open Source ist Open Data ein weiteres Buzzword, das den Zugang zu digitalen Ressourcen thematisiert.

- Public Domain ist die Idee, Programme kostenfrei zu vertreiben. Sie stammt aus der Urzeit der Programmierung. Viele hochwertige Anwendungen sind daraus entstanden.

- Open Source ist der nächste wichtige Meilenstein, denn erst die Freilegung des Quellcodes garantiert, dass Programme transparent werden und viele Entwickler freie Software stetig verbessern können. Mit Linux hat sich das heute erfolgreichste Server-Betriebssystem dieser Idee verschrieben.

- Open Data ist die logische Weiterentwicklung. Vor allem in Verwaltungsorganen, aber auch in öffentlich geförderten Forschungsprojekten oder in Krankenhäusern werden ständig große Datenmengen erzeugt. Diese Daten frei und umfassend zugänglich zu machen, ist die Idee hinter Open Data.

Dieses Konzept ist bestechend: Wenn die Allgemeinheit auf diese mit öffentlichem Geld geförderten Daten vollen Zugriff hat, kann auf vielfältige Weise ein gänzlich neuer Mehrwert entstehen. Aus diesem Ansatz ergeben sich Chancen und Risiken, die wir am Modellierungstag offen diskutieren wollen.

- Grundsätze bei der Publikation von Open Data

- Vorteile und Nachteile von Open Data für Unternehmen

- Qualitätssicherung in Open Data Projekten

- Geschäftsmodelle auf Basis offener Daten

Zu diesen und weiteren Themen erwarten wir interessante Vorträge aus Wissenschaft und Wirtschaft. Auch die öffentliche Verwaltung ist zum Dialog rund um dieses Thema herzlich eingeladen, hängt doch Open Data eng mit Open Government, dem Konzept der offenen Beteiligung der Zivilgesellschaft an der Verwaltung, zusammen.

In the field of robotics, the task of actively balancing on a narrow support, such as a point or line, is usually seen as a control problem; and it has been solved many times using various ideas from control theory. This talk takes a fresh look at the problem from a physical point of view, which yields both a simple model of the physical process of balancing and a simple control system to accomplish it. For the special case of balancing in a plane, the robot’s behaviour is characterized by just two numbers, and the controller’s gains are trivial functions of these numbers. The talk will also explain how it is possible to use a single motion freedom to perform two tasks at once: balance the robot and simultaneously follow a commanded motion trajectory. Several examples of balancing behaviours will be shown and explained.

Due to ever increasing product complexity of today_s and tomorrow’s products, the role of simulation is changing rapidly. Without simulation we would not be able to handle the ever growing complexity of our products at the speed innovations are emerging today. Additionally simulation is not only a key technology for engineering but simulation per se is expected to be of a fundamental value customers are paying for, e.g. as an assist system for optimal operation of products, systems, and infrastructures.

In this talk, I provide a short overview on Computer Aided Engineering in an industrial context as well as highlighting three innovation topics exploiting state-of-the-art mathematical research: Interactive Computer Aided Engineering, Plug & Play Multi-X-Simulation, and Simulation-based Assist Systems.

We start with a brief historical account of inverse problems in imaging. We highlight the emergence of the concept of sparsity, which opened the door to the resolution of more difficult image reconstruction problems, including compressed sensing. We then show the optimality of splines for solving problems with total-variation (TV) regularization constraints, which concludes the deterministic part of our story.
Next, we introduce a statistical formulation in which signals are modeled as sparse stochastic processes. These latter entities (including splines) are solutions of non-Gaussian stochastic differential equations and are intrinsically sparse in the sense that they admit a concise representation in a matched wavelet basis. The formalism is applied to the discretization of ill-conditioned linear inverse problems where both the statistical and physical measurement models are projected onto a linear reconstruction space. This leads to the specification of a general class of maximum a posteriori (MAP) signal estimators complemented with a practical iterative reconstruction scheme. While the framework is backward compatible with the traditional methods of Tikhonov and TV, it opens the door to a much broader class of potential functions that are inherently sparse, while it also suggests alternative Bayesian recovery procedures. The approach is illustrated with the reconstruction of images in a variety of modalities including deconvolution microscopy, phase-contrast tomography, and refractive-index microscopy.

Biography:
Michael Unser is professor and director of EPFL_s Biomedical Imaging Group, Lausanne, Switzerland. His primary area of investigation is biomedical image processing. He is internationally recognized for his research contributions to sampling theory, wavelets, the use of splines for image processing, stochastic processes, and computational bioimaging. He has published over 250 journal papers on those topics. He is the author with P. Tafti of the book "An introduction to sparse stochastic processes", Cambridge University Press 2014.
From 1985 to 1997, he was with the Biomedical Engineering and Instrumentation Program, National Institutes of Health, Bethesda USA, conducting research on bioimaging.
Dr. Unser has held the position of associate Editor-in-Chief (2003-2005) for the IEEE Transactions on Medical Imaging. He is currently member of the editorial boards of SIAM J. Imaging Sciences, IEEE J. Selected Topics in Signal Processing, and Foundations and Trends in Signal Processing. He co-organized the first IEEE International Symposium on Biomedical Imaging (ISBI_2002) and was the founding chair of the technical committee of the IEEE-SP Society on Bio Imaging and Signal Processing (BISP).
Prof. Unser is a fellow of the IEEE (1999), an EURASIP fellow (2009), and a member of the Swiss Academy of Engineering Sciences. He is the recipient of several international prizes including three IEEE-SPS Best Paper Awards and two Technical Achievement Awards from the IEEE (2008 SPS and EMBS 2010).

The formation of stars and their accompanying planets occurs in a dynamic environment in which larger scales couple to smaller scales all the way from the infall of gas from the cosmic web into galaxies down to the accumulation of dust to form planetesimals within protostellar disks. To understand the origins of our own Solar System and the observed populations of stars and planets requires simulations that capture the interactions between these scales while modeling the dominant physical and chemical processes at each scale. (Whether these set the stage for biological processes elsewhere is, of course, one of the great unanswered scientific questions.) In this talk I will review my group’s work on such simulations, using multiple numerical techniques.

Working as a PhD student you have the challenging task of developing research findings and write you doctoral thesis within three years. This alone is a demanding job. In addition, it is vital to the scientific process that your findings are presented to the scientific community. For most PhD students this is the first big project in their professional life and it could have a crucial impact on their future professional career. PhD students are highly motivated when they start their PhD studies but may underestimate the need for professional management for this three-year project \"doctoral thesis\".
This seminar demonstrates how to approach the doctoral thesis in a professional way. Project management tools and techniques are used, tailored to the specific situation of PhD students. You will learn how to set a project vision, define clear objectives, gain buy-in from your supervisor and other colleagues in your group, and how to develop a project plan, which is structured and at the same time flexible enough to easily adjust to unexpected findings. You will establish a \"controlling cycle\" which helps you to recognise risks and problems as early as possible, and you will learn how to manage critical situations and deal with ups and downs. Furthermore, networking with colleagues, supervisors and other people are important topics of this seminar.
Throughout the seminar, you will work on your own doctoral thesis and share your experience with others. This seminar is most beneficial for PhD students who are in the early phases of their doctoral thesis. At the end of the seminar you will have established a strategy on how to approach your own doctoral thesis. During the follow-up REVIEW we will share experience and best practices and deal with open questions from the first module.
This seminar will help you to make the most effective use of your three years and finish your doctoral thesis on time.
You will also learn and practise the basic concepts of project management – a prerequisite in industries and research institutions.

Please register
here.

Ziel des Workshops ist es, vom Lebenslauf über das Vorstellungsgespräch bis zur Frage nach dem Gehalt optimal vorbereitet zu sein.
Bewerben ist wie flirten: Wer zu langweilig und 08/15 ist oder den falschen Ton trifft, wird abgewiesen. Zwei Vorgehensweisen können nun angewandt werden: die quantitative oder die qualitative Strategie.
Die quantitative Strategie bedeutet: mehr schlechte Bewerbungen schreiben. Sprich: einfach mehr flirten in der Hoffnung, dass auch ein blindes Huhn mal ein Korn findet.
Die qualitative Strategie verspricht, durch eine kluge Bewerbung und souveränes Auftreten in den Gesprächen mehr Erfolg zu haben. Das bedeutet beim Bewerben wie beim Flirten: mehr Auswahl. Um letztere Strategie dreht sich das Seminar.

Inhalt:
-Was unterscheidet eine gute von einer schlechten Bewerbung?
-Wie denkt der Personaler?
-Das Vorstellungsgespräch aus beiden Perspektiven beleuchtet.
-Die typischen Stressfragen und die typischen Fehler des Bewerbers.
-Wie viel Gehalt kann ich verlangen?

Nach dem Seminar werden die Teilnehmer:
-Bessere Bewerbungen schreiben können.
-Mehr Einladungen zu Vorstellungsgesprächen haben.
-Im Vorstellungsgespräch punkten und aus der Masse hervorstechen.
-Im Stressinterview unter Druck gelassen bleiben.
-Gekonnt die Gehaltsfrage klären.

Bitte hier anmelden

Discussion of topical challenges to modeling, discretization and adaptive methods and their efficient numerical solution in the context of fluid-structure interaction

After several decades of intensive research, numerical analysis and simulation of fluid-structure interactions remain a challenging topic with a large number of unresolved problems and issues. While the numerical analysis of the coupled system of equations in terms of well-posedness and convergence is typically limited to simple model problems, a lot of insight have been gained over the years by means of numerical simulations. Established methods like the Arbitrary Lagrangian-Eulerian (ALE) method or the Immersed Boundary Method have been succesfully applied to a wide range of applications, including for example aero-elasticity and aero-acoustics, biomechanics, energy or mechanical engineering.

Nevertheless, there are yet a number of problems, where most of the established methods fail or come to a limit. Problems are caused for example by large structural deformations or contact problems, stiff couplings, extreme parameters or a huge computational complexity. In the last years, a number of novel methods and approaches have been developed to tackle such problems many of them being still subject of ongoing research.

An area of research on its own is the development of efficient solvers for the underlying linear systems of equations. The high complexity of real world applications calls for algorithms that include adaptivity in time and space, model reduction, as well as parallelization. In the case of strong couplings, the coupled system of equations is extremely bad conditioned, such that the design of efficient solvers, e.g. multigrid solvers, is a challenge.

This workshop addresses the previously mentioned challenges and aims at bringing together experts and junior scientists in the fields of modeling, adaptive discretizations and solvers for fluid-structure interaction. To provide a platform in order to teach and learn state-of-the art formulations for fluid-structure interaction, this workshop consists of a two-day-school and a subsequent three-day-symposium. The latter one will consist of invited and contributed presentations of junior scientists and experts whereas the school lectures will be given by three young scientists and experts in their field covering each of the three topics of our symposium.

Scalar fields are a versatile and integral part of many applications: In computer graphics, they appear in the form of the classical heat kernel shape descriptors. In molecular chemistry simulations, scalar fields derived from electron density are used to model complex molecular interactions.

Scalar fields are very appealing for analysing and visualizing complex data sets because they permit a precise mathematical theory that is largely based on Morse theory.

In this talk, Julien Tierny of the French National Centre for Scientific Research will give an introduction to this topic and outline past and current research. The talk is specifically geared towards an audience of non-experts in topological methods.

The geometric shortest path problem: computing approximately the shortest path between two points in a geometric domain such as a simple polygon or on a polytope, is a fundamental problem in computational geometry and is already solved using Steiner points and graph tools on the entire domain.

In this talk I will introduce the method of multiple shooting for solving the problem. It includes the factors: (f1) partition of the domain, (f2) the collinear/straightest condition for the shortest path at shooting points, and (f3) refinement of shooting points. In particular, the method no longer relies on Steiner points and graph tools on the entire domain.

Our corresponding algorithms are implemented in C and comparisons with Lee and Preparata`s algorithm (in a simple polygon) on running time, with Agarwal, Har-Peled, and Karia`s algorithm (on a polytope) on the accuracy of the approximate shortest paths, are presented. This is a joint work with N. N. Hai and T. V. Hoai.

Currently, fracture propagation is a major topic in applied mathematics and engineering. It seems to turn out that one of the most promising methods is based on a variational setting and more specifically on a thermodynamically consistent phase-field model. Here a smoothed indicator function determines the crack location and is characterized through a model regularization parameter. In addition, modeling assumes that the fracture can never heal, which is imposed through a temporal constraint, leading to a variational inequality system.
The basic fracture model problem is augmented with several hints and discussions of serious challenges in developing numerical methods for fracture propagation. Key aspects are robust and efficient algorithms for imposing the previously mentioned crack irreversibility constraint, an heuristic solution of the otherwise indefinite Jacobian matrix, computational analysis of the interplay of model and discretization parameters, goal-functional evaluations, coupling to other multiphyics problems such as fluid-filled fractures in porous media, and steps towards high performance computing for tackling practical field problems.

Im Rahmen des Forschungsprojekts "Textdatenbank und Wörterbuch des Klassischen Maya" wird derzeit eine maschinenlesbare Annotation der Hieroglyphentexte auf der Grundlage von XML TEI konzipiert und entworfen. Ziel des strukturierten Markups der Mayaschrift ist es die Struktur der Originalschreibung und die Anordnung der Zeichen im Text möglichst genau mit Hilfe von TEI XML abzbilden. Bei partiell entzifferten Mayaschrift kommt das Problem hinzu, dass anstelle von Transliterationen und Transkriptionen unentzifferte Zeichen oder Textstellen mit Hilfe von Codes und Nomenklaturen wiedergegeben werden müssen um auf dieser Grundlage fragliche Zeichen in ihrem jeweiligen Verwendungskontext zu dokumentieren. Die Dokumentation der Originalschreibung ist grundlegend für die epigraphische Arbeit mit syllabischen und logo-syllabischen Hieroglyphen- und Keilschriftsystemen. Unser Projekt hat hier ein Desiderat für die vergleichende Schriftforschung erkannt und möchte in Zusammenarbeit mit Experten für kompexe Schriftsysteme Lösungen auf der Grundlage von XML TEI und Epidoc erabeiten. Im Rahmen des Vortrags in Heidelberg stellen wir die Probleme und Herausforderungen bei der Annotation der Mayaschrift vor.

Auf Grund der begrenzten Platzkapazität wird um Anmeldung per Doodle gebeten.

Die nur teilweise entzifferte Hieroglyphenschrift und Sprache der Mayakultur steht im Mittelpunkt eines Forschungsprojektes der NRW Akademie der Wissenschaften, das in Kooperation zwischen den Universitäten Bonn und Göttingen durchgeführt wird. Ziel ist die Erstellung einer Textdatenbank und ein darauf basierendes Wörterbuchs des Klassischen Maya. Im Rahmen des Projekts werden die Text- und Bildträger systematisch und nach einheitlichen Standards beschrieben, das Ausgangsmaterial auf der Basis von XML maschinenlesbar gemacht und auf diese Weise die Grundlagen für die Kompilation des Wörterbuchs geschaffen. Dieses Unterfangen wird mit Methoden und Technologien aus den digitalen Geisteswissenschaften in Angriff genommen. Wesentliche Voraussetzung ist dabei, dass nicht nur der sprachliche Inhalt der Inschriften und die ikonischen Informationen der Bilddarstellungen, sondern auch Daten über den Inschriften- und Bilddträger (Beschreibungs- oder Metadaten) berücksichtigt und in einer Datenbank angelegt werden. Zu diesem Zweck werden in TextGrid Tools und Workflows entwickelt, welche die I. Dokumentation der Schrift- und Bildträger mit Aufarbeitung des Forschungsstandes, II. die epigraphisch-linguistische Auswertung der Hieroglyphentexte sowie III. Edition der Texte mit Transliteration, Transkription und Übersetzung in einem einzigen System ermöglichen. Die VRE enthält nicht nur Beschreibungen der Textträger oder Informationen die Texte, sondern der Datenbanknutzer bekommt mit Hilfe der Literaturdatenbank auch einen Überblick darüber, welche Autoren sich mit einem Monument befasst oder es publiziert haben, eine Textpassage diskutiert oder erstmals eine bis heute gültige sprachliche Lesung einer Hieroglyphe präsentiert haben. Der Textträger erhält dadurch eine ‚Biographie‘, die eng mit dem Textinhalten verwoben ist und bei der Bedeutungsanalyse von Wörtern berücksichtigt wird.

Auf Grund der begrenzten Platzkapazität wird um Anmeldung per Doodle gebeten.

Speaker: Prof. Branislav Borovac, Mechatronics, Robotics and Automation, University of Novi Sad, Serbia

This talk is part of the ORB Oberseminar that for this week has been shifted to: Tuesday, Feb 16, 2016 at 15:00 at Speyerer Str. 6, room H2.22

Traffic optimization is a big field related to many subjects, such as, mathematics, computer science, or physic. In big cities, people are suffering from traffic problems, for example traffic jams, traffic accidents. Thus, improving traffic systems in urban areas becomes more and more important. The 1-day workshop on "Traffic Optimization" is organized by the research group for Discrete and Combinatorial Optimization on 8th October, 2015. The workshop aims at addressing the many aspects of traffic optimization, in particular in urban environments. The main purpose is to introduce the field to general audience. There are about 6-8 experts coming from Germany, Italy, Hungary, and Spain. Each one will give a 45-minutes talk about some topics as followings:

- Routing problems, such as shortest path, dissimilar paths, etc.
- Traffic assignment modeling
- Traffic simulation

More information are available in the webpage.

Probabilistic graphical models provide a consistent framework for the statistical modeling and the computational analysis of scientific empirical data. The past decade has witnessed a significant increase in respective research in the field of image analysis and related application areas, driven by the synergy between statistics, pattern recognition, computer vision and machine learning. The objective is to devise models that enable to infer a coherent global interpretation of noisy and ambiguous local image measurements, taking into account spatiotemporal context in images and videos, and domain-specific contextual knowledge.

Applications of probabilistic graphical models to such large-scale problems raise numerous research problems of modeling and algorithm design for inference and learning, requiring interdisciplinary expertise in applied mathematics, computer science and physics, besides a profound knowledge of the respective application areas.

The basic intention of the Research Training Group is to gather experts from these fields and to establish a coherent research and study program on probabilistic graphical models, with a focus on spatial and spatiotemporal models and their applications in image analysis. The project treats methodological basic research on an equal footing with challenging scientific applications of image analysis in environmental science, life sciences and industry.

The Research Training Group will provide a scientifically unique environment for study, collaboration and innovative research on probabilistic graphical models across disciplines, producing highly-qualified candidates for research careers in academia and industry.

Confirmed speakers:

Mario A. T. Figueiredo, Instituto Superior Técnico, Lisboa, Portugal
Victor Lempitsky, Skoltech Computer Vision, Moskau, Russia
Raquel Urtasun, University of Toronto, Toronto, Canada
Max Welling, University of Amsterdam, Amsterdam, Netherlands

Please register free of charge here by September 30, 2015.

Please register here

Manche Studien behaupten: Der Steuerspartrieb der Deutschen ist stärker ausgeprägt als der Sexualtrieb. Nun haben Studenten im Studium nur selten Steuerabgaben zu leisten, aber dafür eine Vielzahl an Studienkosten (z.B. für den Laptop, für Lehrmaterialien, Fahrtkosten,Telefon, Miete u.v.m.). Und diese Ausgaben können vom Finanzamt zurückgeholt werden, spätestens ab dem Zeitpunkt in dem Steuern anfallen. Erfahrungsgemäß bedeutet das im ersten Berufsjahr eine Steuererstattung in 4-stelliger Höhe.

Darüber hinaus gibt es auch wirtschaftliche Themen, die im Studium / Promotion bewegt werden sollten und die eine immer wiederkehrende Steuererstattung ermöglichen. Dies wird ebenfalls aufgezeigt.

Nach dem Training werden die Teilnehmer:
-die Vorgehensweise kennen, wie sie ihre Studienkosten rückerstattet bekommen.
-für sich wichtige Themen erkennen und einen der größten Steuerhebel nutzen können.
-wissen, wie sie einen Verlustvortrag generieren und Werbungskosten und Sonderausgaben steueroptimiert einordnen können.

Anmeldung bitte hier