HGS MathComp Curriculum & Events

Der Datenbank- und Softwareriese SAS öffnet seine Türen für Mathematikerinnen und Frauen* aus den STEM-Feldern! Euch erwarten Vorträge und Hands-On-Erfahrungen mit den SAS-Toolsets sowie die Möglichkeit zur Vernetzung mit dem Women´s Initiative Network von SAS.

Zielpublikum:
Diese Exkursion ist Teil des Upstream Programms und nur für weibliche Mitglieder zugänglich.

! ANMELDUNG ERFORDERLICH !
Anmeldungen für den Workshop bitte an: upstream@iwr.uni-heidelberg.de

Deep learning-based approaches have revolutionized virtually all domains of computer vision, including the field of microscopy image analysis. In image reconstruction and classification, in segmentation and artificial labelling, they have pushed both flagship projects and bread-and-butter everyday tasks, allowing image analysis to keep pace with the recent advancements in imaging technology and instrumentation. I will talk about the recent work of my group that has enabled the first segmentation of cells in a whole animal imaged with electron microscopy, our collaboration with microscope developers for trusted reconstruction of light field microscopy and our current efforts to reduce the annotation budget for training of segmentation algorithms. I will also show how we strive to make our methods accessible to biologists without computational expertise through our software ilastik and through the emerging community network collection at the BioImage Model Zoo.

The IWR Colloquium will be held as an in-person event at the Mathematikon. In addition it will be streamed via Zoom.
For more information please visit the website of the colloquium.

Link: www.iwr.uni-heidelberg.de/events/iwr-colloquium

HGS MathComp Members will receive 1 ECTS credit for every 4 talks attended. Please make sure to include them in your BlueSheet.

Predicting the emergent properties of a material from a microscopic description is a scientific challenge. Machine learning and reverse-engineering have opened new paradigms in the understanding and design of materials. However, the soft-matter field has lagged far behind in embracing this approach for materials design. The main difficulty stems from the importance of entropy, the ubiquity of multi-scale and many-body interactions, and the prevalence of non-equilibrium and active matter systems. The abundance of exotic soft-matter phases with (partial) orientation and positional order like liquid crystals, quasicrystals, plastic crystals, along with the omnipresent thermal noise, makes the classification of these states of matter using ML tools highly non-trivial. In this talk, I will address questions like: Can we use machine learning to autonomously identify local structures, detect phase transitions, classify phases and find the corresponding order parameters in soft-matter systems, can we identify the kinetic pathways for phase transformations, and can we use machine learning to coarse-grain our models? Finally, I will show how one can use machine learning to reverse-engineer the particle interactions to stabilize nature’s impossible phase of matter, namely quasicrystals?

The IWR Colloquium will be held as an in-person event at the Mathematikon. In addition it will be streamed via Zoom.
For more information please visit the website of the colloquium.

Link: www.iwr.uni-heidelberg.de/events/iwr-colloquium

HGS MathComp Members will receive 1 ECTS credit for every 4 talks attended. Please make sure to include them in your BlueSheet.

In statistics, stochastic processes are the main tool for modelling spatial and temporal data, and the most commonly used type are the Gaussian processes. A popular approach to define Gaussian processes is through linear stochastic differential equations driven by white noise. This methodology is often denoted the SPDE approach. By utilising methods from numerical analysis one has been able solve several of the computational bottlenecks that have hampered the usage of Gaussian processes for real world data sets, in particular for spatial statistics. I will focus on how one can apply the same methodology to create non-Gaussian processes, in space and time. I will also discuss what practical properties these processes have over their Gaussian counterpart. Further, I will present very recent work where we have extended the SPDE approach for Gaussian process, to Euclidean graphs (like street networks). Here the linear differential operators are obtained using Quantum graphs, which opens these methods for a new wide range of applications.

The IWR Colloquium will be held as an in-person event at the Mathematikon. In addition it will be streamed via Zoom.
For more information please visit the website of the colloquium.

Link: www.iwr.uni-heidelberg.de/events/iwr-colloquium

HGS MathComp Members will receive 1 ECTS credit for every 4 talks attended. Please make sure to include them in your BlueSheet.

In this talk I will present a geometric multilevel optimization approach choosing as case study a regularised inverse problem. In particular, the approach is motivated by variational models that arise as the discretization of some underlying infinite dimensional problem. Such problems naturally lead to a hierarchy of discretized models. We employ multilevel optimization to take advantage of this hierarchy: while working at the fine level we compute the search direction based on a coarse model. By utilising concepts of information geometry in our formulation, we propose a smoothing operator that only uses first-order information and incorporates constraints smoothly. We show that the proposed algorithm is well suited for ill-posed reconstruction problems and demonstrate its efficiency on several large-scale examples.

The IWR Colloquium will be held as an in-person event at the Mathematikon. In addition it will be streamed via Zoom.
For more information please visit the website of the colloquium.

Link: www.iwr.uni-heidelberg.de/events/iwr-colloquium

HGS MathComp Members will receive 1 ECTS credit for every 4 talks attended. Please make sure to include them in your BlueSheet.

From the vision “Stability, Prosperity and Sustainability,” Thailand has adopted “Thailand 4.0” as a policy vision for economic development which aims at driving the country out of a middle-income towards a high-income economy. Underlying this is no other than an attempt to change from the current economic model towards an “innovation-driven economy”—i.e., transitioning from an industry-driven development to a technology-, creativity-, and innovation-driven development by consolidating as well as integrating various fields of knowledge. This, thus, necessitates deepening mathematical research on optimization and numerical methods as it plays a considerable role in providing solutions for a variety of problems including those in such areas as industry, finance, engineering, economics and agriculture.

Realizing the importance of research on optimization and numerical methods, the Department of Mathematics, Faculty of Science, King Mongkut’s University of Technology Thonburi—in cooperation with the Interdisciplinary Center for Scientific Computing (IWR), Heidelberg University, the Federal Republic of Germany, and along with the nationwide as well as the Southeast Asian regionwide network of universities including Walailak University, Prince of Songkla University (Pattani Campus), Chulalongkorn University, Silpakorn University, Mahidol University, Chiang Mai University, and King Mongkut’s University of Technology North Bangkok—will organize an international “Workshop on Computational Science 2022: Webinar on Optimization and Numerical Methods in Industries.”

This year, the Department will co-host the event with the Department of Mathematics, School of Science, Walailak University, and the Department of Mathematics and Computer Science, Faculty of Science, Chulalongkorn University.

Topic:
Natural or engineered materials often contain two or more key constituents, arranged in a heterogeneous structure varying at different scales. Such materials are desirable because their macroscopic properties can be superior to the properties of the individual constituents. It is even possible to explicitly design them for a particular purpose by changing the composition of the constituents.

An example are carbon fibre composites for lightweight structures and vehicles. The mathematical modelling of such heterogeneous or composite materials typically leads to PDEs with highly oscillating coefficients. Direct numerical solution of such problems with traditional methods, such as finite elements is computationally expensive. Just to compute the correct qualitative behaviour, the mesh resolution would need to be sufficiently high to capture all the fine scale variation.

In this seminar, we will study multiscale numerical methods that address various aspects of this challenge. This includes the generation of low-dimensional yet high-quality approximation spaces, acceleration of fine-scale solutions and efficient uncertainty quantification for multiscale problems.

Each student will present a key publication on multiscale numerical methods with an aim to cover most aspects of the field.

Prerequisites:
Basic knowledge of partial differential equations, Sobolev spaces and Finite Element methods is required. For the uncertainty quantification topics, basics of probability theory are needed.

Registration and Schedule:
• First meeting: April 22, 2022, 13:00. Room: TBA
• Since we use MÜSLI for email communication, please register at: https://muesli.mathi.uni-heidelberg.de/lecture/view/1517
• Schedule: Will be chosen on first meeting to maximize attendance

Contact:
Main contact - Linus Seelinger: linus.seelinger at iwr.uni-heidelberg.de
Secondary contact - Robert Scheichl: r.scheichl at uni-heidelberg.de

Talks:
The length of each talk is 40 min. + 10 min. for questions and discussion.
Before your talk you should meet with one of us to discuss your presentation.

Prof. Morteza Fotouhi • Sharif University of Technology

Dates:
April 20, 2022 / 09:00-12:00
April 21, 2022 / 14:00-17:00

The term moving interfaces and free boundary refers to a class of problems in which the domain of the problem itself is also an a priori unknown, as well as the basic unknown solution to governing equations. Hence, finding the domain is part of the problem.
In this lecture I shall present some basic models in moving interfaces and free boundary problems. These includes obstacle problem, Stefan problem, Hele-Shaw flow and Muskat problem.
I will also review the state of art for obstacle problem. The lectures will be at elementary level for both advanced master and Ph.D. students.

! REGISTRATION REQUIRED !
Please register via the office of the graduate school: hgs@iwr.uni-heidelberg.de

Have you sat in front of an empty sheet of paper or an empty file before, trying to start a report, a paper or your PhD Thesis, but struggling to make progress? Or have you had to read a research article relevant to your PhD project that was so incomprehensible or dry that you were wishing to be somewhere else for the day to end? Good scientific writing is an absolutely essential part of our work as researchers; only through good communication will we be able to captivate the audience and truly enthuse them for our achievements and results. It is as important as pure academic ability for a successful professional career in academia, but also in industry, but typically we are not trained formally in good writing in a mathematics or science degree. In this crash course I aim to give you some useful pointers on good writing in the mathematical sciences, and through some hands-on exercises we will try to immediately put them in practice.

! REGISTRATION REQUIRED !
Please register via the office of the graduate school: hgs@iwr.uni-heidelberg.de

Netzwerken ist ein wesentlicher Bestandteil bei der Entwicklung einer wissenschaftlichen Karriere. Die Analyse, der Aufbau und die Pflege von (wissenschaftlichen) Netzwerken unterstützen auf dem Weg zur Professur aber auch bei außeruniversitären Karrierezielen. Dabei gilt es die eigene komplexe Forschungstätigkeit in sich bietenden Situationen innerhalb (z.B. wissenschaftliche Gemeinschaft, Konferenz) und außerhalb (z.B.: Fördereinrichtungen, Wirtschaft) der Wissenschaft überzeugend und verständlich darzustellen.

In diesem Workshop geht es darum, sich auf Netzwerksituationen wie bspw. den Besuch einer wissenschaftlichen Konferenz vor- und nachzubereiten. Dazu werden verschiedene Situationen vorgestellt und in Übungen und Rollenspielen mit Videoaufzeichnungen (falls gewünscht) ausprobiert und im Detail besprochen.

Des Weiteren wird das eigene Netzwerk analysiert, um Potentiale für dessen Weiterentwicklung zu erkennen und diese bei entsprechenden Netzwerkgelegenheiten zu nutzen. Überdies werden die besten Tricks und Kniffe für erfolgreiches Smalltalken vermittelt, um die eigene Kommunikation in Alltag und Beruf zu optimieren.

Themenauswahl:
• Netzwerkpotenziale: Was ist Netzwerken und warum brauche ich es?
• Netzwerkanalyse im Kontext Wissenschaft – die Rolle von Mentoren
• Analyse, Aufbau und Pflege von (wissenschaftlichen) Netzwerken – wie sieht mein eigenes Netzwerk aus?
• Methodentools: Elevator Pitch zur (überzeugenden und verständlichen) Darstellung von komplexen Forschungstätigkeiten
• Methodentool: „Conference Dinner/Conference Break“
• Methodentool: „Erfolgreiche Selbst-Präsentation“: Story Telling und Grundregeln für erfolgreiches Smalltalken

Zielpublikum:
Dieser Workshop ist Teil des Upstream Programms und nur für weibliche Mitglieder zugänglich.

! ANMELDUNG ERFORDERLICH !
Anmeldungen für den Workshop bitte an: upstream@iwr.uni-heidelberg.de

Datum:
10.05.2022 und 12.05.2022 (14:00-16:30 Uhr)

Beschreibung:
Vielfalt und Pluralität sind mittlerweile Alltag in unserer Gesellschaft, im Privat- und Berufsleben: Menschen mit unterschiedlichsten Lebensentwürfen, Hintergründen, Glauben oder Herkunft treffen aufeinander.

In dem zweiteiligen Workshop werden wir uns mit verschiedenen theoretischen Ansätzen im Themenfeld Diskriminierung beschäftigen und anhand praktischer Übungen einen eigenen erfahrungsbasierten Zugang dazu finden. Ziel ist es, Selbstverständlichkeiten und gängige Narrative diskriminierungskritisch zu hinterfragen sowie Handlungsoptionen aufzuzeigen, die tradierte Verhaltensmuster aufbrechen können. Ein besonderes Augenmerk soll dabei auf die Geschlechtergleichberechtigung gelegt werden.

Dozentinnen:
Halszka ?liwa-Ohnesorge verantwortet die Bildungsstelle Plurales Heidelberg bei Mosaik Deutschland e.V. Sie studierte Religionswissenschaft und Politikwissenschaft Südasiens an der Universität Heidelberg. Nachdem Studium arbeitete sie u.a. für ein ziviles, indisch-pakistanisches Friedensprojekt, bei dem die gleichberechtigte Förderung von Jungen und Mädchen ein zentrales Element war. Heute liegt der Schwerpunkt ihrer Arbeit auf der vorurteilsbewussten und diskriminierungskritischen politischen Bildung.

Lara Track leitet das Antidiskriminierungsbüro Heidelberg (ADB) bei Mosaik Deutschland e. V. Sie berät Menschen, die Diskriminierung erfahren haben und setzt sich in Netzwerken auf kommunaler, Landes- und Bundesebene gegen Diskriminierung ein. Im Rahmen ihrer wissenschaftlichen Mitarbeit am Historischen Seminar der Universität Heidelberg erforscht sie Verbindungen zwischen Frauenfriedensaktivismus und Feminismus in den USA. In der Auseinandersetzung mit US-Feminismen gründet ihr Interesse am Konzept der Intersektionalität, das in der mehrdimensionalen Beratung im ADB zum Tragen kommt.

Zielpublikum:
Mitarbeiterinnen und Mitarbeiter in der Verwaltung, PostDocs, Junior Forschungsruppenleiterinnen und -leiter

! ANMELDUNG ERFORDERLICH !
Anmeldungen für den Workshop bitte an: hgs@iwr.uni-heidelberg.de
[Anmeldeschluss: 18.04.2022]

The graduate school HGS MathComp will organize an Integrative Think Tank (ITT) from July 4-8, 2022 in Heidelberg.

What is an "Integrative Think Tank" (ITT)?
ITT is a 1-week-challenge workshop with two industrial partners. Students and lecturers work together with experts from the companies to investigate R&D topics related to current developments at these companies.

What is the goal of the event?
The goal of the event is to formulate actual research topics as thesis projects outlines (master or PhD). After the event the University and the partner companies join forces to convert the project outlines into actual research projects.

Who can participate?
The event is open to all master and PhD students at Heidelberg University. ITT is most successful if different students from various fields work together to analyze the research questions of our partners.

What is the reward?

• All participants get a certificate of attendence that certifies 2 ECTS as key competence training.
• Attending students can also apply to participate in the projects defined at ITT as their master or PhD project.
• On top you get direct contacts to the participating companies for internships and practicals (or later employment ...).

OK, you got me! How do I register?
Scroll to the end of the page and find all details regarding registration. Get some more information along the way! And: Tell your friends about it! The more, the merrier!

For more information please visit the website:

www.mathcomp.uni-heidelberg.de/itt-2022

Scientific progress in machine learning is driven by empirical studies that evaluate the relative quality of models. The goal of such an evaluation is to compare machine learning methods themselves, not to reproduce single test-set evaluations of particular optimized instances of trained models. The practice of reporting performance scores of single best models is particularly inadequate for deep learning because of a strong dependence of their performance on various sources of randomness. Such an evaluation practice raises methodological questions of whether a model predicts what it purports to predict (validity), whether a models performance is consistent across replications of the training process (reliability), and whether a performance difference between two models is due to chance (significance). The goal of this tutorial is to provide answers to these questions by concrete statistical tests.

The tutorial is hands-on and accompanied by a textbook (Riezler & Hagmann: Validity, Reliability, and Significance: Empirical Methods for NLP and Data Science Synthesis Lectures on Human Language Technologies, Morgan & Claypool Publishers, 2022) and a webpage including R and Python code: https://www.cl.uni-heidelberg.de/statnlpgroup/empirical_methods/

! REGISTRATION REQUIRED !
Please register via the office of the graduate school: hgs@iwr.uni-heidelberg.de