[External Email] Subject: NA Digest, V. 20, # 43 NA Digest Monday, November 09, 2020 Volume 20 : Issue 43 Today's Editor: Daniel M. Dunlavy Sandia National Labs [log in to unmask] Today's Topics: New Book, Simplicial Partitions with Applications to the FEM AI and IoT for Flow Modeling, ONLINE, Nov 2020 Senior-Level Position, Mathematical Analysis, McMaster Univ Postdoc Position, Computational Fluid Mechanics, Brazil Postdoc Position, Numerical Optimization, Argonne NL Postdoc Position, TU Berlin, Germany Postdoc Position, Theoretical Fluid Mechanics, McMaster Postdoc Positions, Applied Mathematics, Columbia Univ Postdoc Positions, Comp Mechanics, Czech Technical Univ Special Issue on Domain Decomposition Methods Special Issue, NA with Applications in Machine Learning Subscribe, unsubscribe, change address, or for na-digest archives: http://www.netlib.org/na-digest-html/faq.html Submissions for NA Digest: http://icl.utk.edu/na-digest/ ------------------------------------------------------- From: Sergey Korotov [log in to unmask] Date: November 05, 2020 Subject: New Book, Simplicial Partitions with Applications to the FEM Simplicial Partitions with Applications to the Finite Element Method (Springer Monographs in Mathematics) 1st ed. 2020 Edition by Jan Brandts, Sergey Korotov, Michal Krizek. This monograph focuses on the mathematical and numerical analysis of simplicial partitions and the finite element method. This active area of research has become an essential part of physics and engineering, for example in the study of problems involving heat conduction, linear elasticity, semiconductors, Maxwell's equations, Einstein's equations and magnetic and gravitational fields. These problems require the simulation of various phenomena and physical fields over complicated structures in three (and higher) dimensions. Since not all structures can be decomposed into simpler objects like d- dimensional rectangular blocks, simplicial partitions are important. In this book an emphasis is placed on angle conditions guaranteeing the convergence of the finite element method for elliptic PDEs with given boundary conditions. It is aimed at a general mathematical audience who is assumed to be familiar with only a few basic results from linear algebra, geometry, and mathematical and numerical analysis. The link to the table of contents is at: https://www.springer.com/gp/book/9783030556761 ------------------------------------------------------- From: Kees Oosterlee [log in to unmask] Date: November 09, 2020 Subject: AI and IoT for Flow Modeling, ONLINE, Nov 2020 AI and IoT for Flow Modeling Friday 20th November 2020 The workshop is organized as part of the Indo-Dutch project, "Digital Twins for pipeline transport networks". The aim of the project is to develop a digital twin that connects sensor data and advanced fluid solvers in order to detect possible leakage of fluid from the pipeline in real-time. Of particular interest then is also the development of AI based fluid flow solvers, as traditional fluid flow models are typically much too slow for real-time applications. We thank the NWO (the Netherlands), MeiTY (India) and Shell (the Netherlands) for funding the project. As part of the workshop the following talks have been scheduled: - 11:00 Jan S. Hesthaven (EPFL): Digital Twins at the interface between modeling, measurements, and machine learning - 11:45 Yogesh Simhan, (IISc Bangalore): IoT and Analytics for Social Good - 12:30 Vineet Tyagi (IIT Bombay): Neural Networks for predicting flow parameters in a pipeline network - 13:00 Amritendu Mukherjee (IISc Bangalore): A comprehensive study to understand the relationship of urbanization and population density with GRACE TWS for selected study regions in India during 2003-2017 - 13:30 Nikolaj T. Mucke (CWI): Reduced Order Modeling for Fluid Simulations using Deep Learning - 14:00 Ruud Henkes (Shell): The role of simulation in leak detection for pipeline operations This workshop will take place, on-line, via zoom.us. For the log-in details see https://www.cwi.nl/research/groups/scientific-computing/events/workshop-20- november-2020/ai-and-iot-for-flow-modeling ------------------------------------------------------- From: Bartosz Protas [log in to unmask] Date: November 05, 2020 Subject: Senior-Level Position, Mathematical Analysis, McMaster Univ There is a senior-level opening in the broad area of mathematical analysis and applications (which also includes numerical analysis) in the Department of Mathematics and Statistics at McMaster University. More information about this position is available at the following webpage https://www.mathjobs.org/jobs/application/16437 The expected start date for this position is July 1, 2021, and the application deadline is December 1, 2020. ------------------------------------------------------- From: João Luiz F. Azevedo [log in to unmask] Date: November 05, 2020 Subject: Postdoc Position, Computational Fluid Mechanics, Brazil Instituto de Aeronautica e Espaco (IAE), at the city of Sao Jose dos Campos, Sao Paulo, Brazil, welcomes applications for a post-doc fellowship in the project "Aerothermodynamic Analysis of Hypersonic Flows with Applications in Atmospheric Reentry Procedure". The fellowship is funded by the Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP). The accepted candidate will receive a monthly stipend of R$ 7.373,10 for 12 months, with possibility of extension for another 12 months. The post-docs are allowed a grant, in the amount of 15% of the annual fellowship, to finance items related to the research activities. The present project concerns numerical simulations of hypersonic reactive flows in thermal non-equilibrium. Main objectives of the work to be performed are twofold: to address the effects of bulk viscosity in hypersonic simulations, particularly for Mars entry-type flows, and to develop the capability to simulate hypersonic flows involving coupled radiative and convective heat transfer mechanisms. The position is initially offered for 1 year with possible reappointment. The researcher will work under the supervision of Prof. Azevedo at IAE and ITA, in Sao Jose dos Campos, Brazil. Applicants must have received a doctorate in Applied Mathematics, Mechanical or Aerospace Engineering, or related discipline, within the last 5 years. Selection criteria require demonstrated research ability in CFD, with a strong background in the physical, mathematical and computational aspects of numerical simulation of hypersonic flows, commitment to collaborative research, and excellent verbal and written skills. To apply, please e-mail (only PDF files) the following items to [log in to unmask] until 4 Dec. 2020: 1. Letter of interest, with full contact information and citizenship/immigration status, concisely addressing the research themes above; 2. Full CV ; 3. Cover letter with names/contact information for 3 references (no letters please). Indicate "CEPID Postdoc" in the subject field. ------------------------------------------------------- From: Jeffrey Larson [log in to unmask] Date: November 04, 2020 Subject: Postdoc Position, Numerical Optimization, Argonne NL The Mathematics and Computer Science Division at Argonne National Laboratory is seeking postdoctoral candidates with expertise in numerical optimization to help solve important quantum information science problems. Applicants with interest in theory, algorithms, and/or software are encouraged to apply. We especially encourage applications from researchers with interests in stochastic optimization, optimal control, derivative-free optimization, or combinatorial optimization. U.S. citizenship is not required. Candidates must have a recent doctoral degree in applied mathematics, statistics, computer science, industrial/systems engineering, or related field. Expertise in one or more areas of numerical optimization is also required. This position is available immediately, but there is flexibility in start dates. Applications received by December 15, 2020 will receive full consideration. Questions can be addressed to Jeffrey Larson ([log in to unmask]) or Sven Leyffer ([log in to unmask]). For more information, please see: https://bit.ly/3jW6dKc ------------------------------------------------------- From: Tobias Breiten [log in to unmask] Date: November 05, 2020 Subject: Postdoc Position, TU Berlin, Germany The Technical University of Berlin, Institute of Mathematics, invites applications for a Postdoc (m-f-d) for a period of max. 2 years. The research is part of the project "Optimal control of stochastic modified equations for the efficient parametrisation of deep neural networks" within the framework of the Cluster of Excellence MATH+, in collaboration with the BTU Cottbus-Senftenberg (Prof. Dr. Carsten Hartmann). Requirements are successfully completed university degree (Master, Diploma or equivalent) as well as PhD in mathematics or related area; detailed knowledge in the areas of control theory, optimization with stochastic and/or partial differential equations as well as numerical methods; preliminary work on Fokker-Planck or Langevin equations is desirable; good command of German and/or English, both written and spoken, willingness to acquire lacking language skills. Application deadline: November 20th, 2020. For the full advertisement, please go to https://stellenticket.de/84926/?lang=en For further information, please contact Tobias Breiten, tobias.breiten (at) tu-berlin.de ------------------------------------------------------- From: Bartosz Protas [log in to unmask] Date: November 04, 2020 Subject: Postdoc Position, Theoretical Fluid Mechanics, McMaster An opening for a post-doctoral fellow is anticipated in Dr. Protas' research group at McMaster University with a start date of September 1, 2021. The focus of this position will be fundamental investigations of extreme behavior, such as potential singularity formation, in fluid flow models using a combination of mathematical analysis and large-scale computations. Expected background involves (ideally, a combination of) theoretical fluid mechanics, large-scale scientific computing (including numerical optimization) and PDE analysis. The position will also involve a limited amount of teaching. Applications should be submitted via www.mathjobs.org (Fellowship ID: McMaster-PDF [#16611]; the advertisement on www.mathjobs.org contains further details concerning the opening and the application procedure). Interested candidates may also contact Dr. Bartosz Protas Department of Mathematics & Statistics McMaster University Hamilton, Ontario, CANADA L8S 4K1 Email: [log in to unmask] URL: http://www.math.mcmaster.ca/bprotas for additional information. ------------------------------------------------------- From: Kui Ren [log in to unmask] Date: November 03, 2020 Subject: Postdoc Positions, Applied Mathematics, Columbia Univ The Program in Applied Mathematics (https://appliedmath.apam.columbia.edu) at Columbia University invites applications for two postdoctoral positions in applied mathematics (broadly defined) starting Fall 2021. The positions are partially supported by an Applied Mathematics Research Training Group (RTG) grant from the National Science Foundation. The postdoctoral positions are appointed annually and can be renewed up to a total of three years. The appointments include teaching duties with an expected teaching load of one course per semester. Applications should be submitted through mathjobs.org at: https://www.mathjobs.org/jobs/list/16622 For full consideration applicants are encouraged to submit all materials before December 1, 2020. Following the rules of the National Science Foundation, the postdoctoral positions are restricted to US citizens, nationals, and permanent residents. ------------------------------------------------------- From: Jan Zeman [log in to unmask] Date: November 05, 2020 Subject: Postdoc Positions, Comp Mechanics, Czech Technical Univ Two fully-funded postdoctoral positions in computational mechanics of materials and structures are available at the Department of Mechanics, Faculty of Civil Engineering, Czech Technical University in Prague. Please consult https://euraxess.ec.europa.eu/jobs/573988 for full opening details. The application deadline is 5 December 2020. ------------------------------------------------------- From: Victorita Dolean [log in to unmask] Date: November 03, 2020 Subject: Special Issue on Domain Decomposition Methods We would like to draw your attention to a special issue on Domain Decomposition Methods https://www.mdpi.com/journal/mca/special_issues/DDM Mathematical modelling in science and engineering problems relies heavily on partial differential equations. Accurate discretization of such PDEs is very often required, and this usually leads to potentially very large linear systems that must be solved in parallel. The computational resources (in terms of hardware) and computational time available can limit the high- fidelity of these simulations. With the advent of parallel computers and the availability of large computational clusters, algorithmic improvements are key in reducing the computational time and increasing the model complexity and accuracy. One of the success stories of parallel computing is linear solvers, but also hybrid solvers, like domain decomposition methods. Contributions related to the development and analysis of the domain decomposition solvers with their different aspects (linear or non-linear, multilevel methods, scalability, HPC implementation, coupling of mathematical models, and computational challenges of large-scale problems) are welcome in this Special Issue. Authors are also invited to submit any other relevant complementary materials, such as software or available links illustrating their research. ------------------------------------------------------- From: Panagiota Tsompanopoulou [log in to unmask] Date: November 03, 2020 Subject: Special Issue, NA with Applications in Machine Learning We would like to draw your attention to a special issue on Numerical Analysis with Applications in Machine Learning: https://www.mdpi.com/journal/mathematics/special_issues/Nume_Analy_MacLear The collection of large amounts of data produced by an enormous variety of users has been a fact for many years now. Therefore, there is a tremendous need to study, analyze, and process these data in order to clear out the possible noise and derive the substantial information. The methods used for such problems constitute the scientific area of machine learning. Looking closer to the theory that supports these methods, one recognizes many fields of numerical analysis, such as Euclidean spaces with metrics and norms, approximation theory, optimization theory, theory of matrices, etc. The use of existing methods and the tuning of their parameters still gives very interesting results for the treated problems. Wishing to go further in the solution of already met or new and more difficult problems, scientists have to go back to the roots of the mathematics used in machine learning, to study, research and create new, stable, and accurate methods. Through this Special Issue, we invite our colleagues to submit articles that rely on numerical analysis methods to address problems in the field of machine learning, presenting both theoretical and experimental results. The fields of interest originate from mathematics and computer science, including (but not limited to) numerical linear algebra, Euclidean, pseudo- Euclidean and metric spaces, theory of matrices, theory of approximation and optimization, machine learning, computer vision, classification, clustering, and pattern recognition. ------------------------------------------------------- End of Digest **************************