Our MINT-EC Math-Talent-School 2022

A Week at the Fraunhofer ITWM to Get a Taste of the Professional World of Mathematics

What does the professional world of a mathematician look like and what is applied mathematics? That's what 15 female students from schools in the national excellence school network MINT-EC can find out at our institute. Our second Math Talent School 2022 from October 10 to 14 is organized by MINT-EC in cooperation with the Felix Klein Center for Mathematics.

The students of grades 11-13 work in teams on different problems with the help of mathematical modeling and computer simulations. And in doing so, they are absolutely motivated!

»Learning together is the best way to form the ultimate unit that can solve any puzzle.«

This quote, for example, comes from the motivation letter of student Anna Hohleweg. In this and similar ways, all the participants express their enthusiasm for working in joint mathematics projects, three of which are available for them to choose from:

  • Rocket launches with computer solutions
  • Crimes in crisis situations
  • Artificial Intelligence and Reinforcement Learning

The results of the work are processed, presented at the end of the Math-Talent-School and discussed together. In addition, the participants visit our institute as well as the Department of Mathematics at the TU Kaiserslautern, where they also receive information about studying mathematics.

On this page we collect impressions, statements, interviews, articles and photos of the MINT-EC-Math-Talent-School 2022.

Project Groups

The three project topics are thematically as diverse as they are interesting – and the best part is: they all tie in with current research areas within applied mathematics!

No Rocket-Science?

Who wants to go to the beach during the vacations these days? Nobody! The new trend is vacationing in space and on the moon – and this is where our experts in mathematics and physics come in. Everything has to be right when the rocket is launched so that it follows the right trajectory and reaches the moon safely.

 

Such Things Can Hardly Be Calculated on Paper

On the way to a computer solution for this problem, the team will work on the following intermediate questions:

  • Suppose we were to launch our tourist at Earth with a large slingshot, can we calculate on paper how (high) it will fly?
  • How can we teach a computer to do that for us?
  • Can we also calculate the trajectory of the rocket this way, if we include the thrust of the weight loss from burning fuel and the drag in our equations?

A simple mathematical procedure for this is derived and translated into the programming language »Python« to solve the relatively difficult »rocket equation« and land safely on the moon.

The group is supervised by our colleagues Armin Bosten and Tobias Ruhwedel from the division »Mathematics for Vehicle Engineering« at Fraunhofer ITWM.

No Rocket Science?
© Fraunhofer ITWM
With the help of their two supervisors, the Rocket Group is sure to land safely on the moon in the end.
No Rocket Science? Tobias Ruhwedel
© Fraunhofer ITWM
One of the two supervisors of the group is ITWM PhD student Tobias Ruhwedel.
Rocket Group: Armin Bosten Supports a Participant.
© Fraunhofer ITWM
Armin Bosten Supports a Participant.

How Do Criminal Offenses Change in »Crisis« Situations?

Since 2012, the German Federal Criminal Police Office has published a nationwide overview of the crimes recorded by the police in the police crime statistics (PKS) once a year.

Now, there have been different so-called crisis situations in the last ten years. Examples of this are the large migration movements as a result of the Syrian war in 2014/15 or the Corona pandemic, which repeatedly paralyzes and restricts parts of public life. With Brexit, the UK's exit from the EU, some also speak of a crisis.
 

Mathematics is Called Upon in Such Contexts!

The team will investigate what impact such events have on the crime situation in Germany:

  • For example, do politically motivated crimes increase in »crisis« situations? 
  • Did the number of home burglaries drop during the corona pandemic and did the number of domestic violence crimes increase? 
  • To what extent does the number of crimes change and when do the numbers return to baseline?

The team will grapple with these questions and more at the level of data analysis and work to model future crises in advance. The goal is to make predictions for various crisis situations that are accurate enough for the police to use in their investigations.

The group is supervised by Paul Weber from the Department of Mathematics (AG Optimization) at the TU Kaiserslautern.

How do Criminal Offenses Change in »Crisis« Situations?
© Fraunhofer ITWM
The participants are particularly concerned with data analysis and how can future crises be modeled in advance?
Group Artificial Intelligence in the Context of Reinforcement Learning
© Fraunhofer ITWM
The team investigates the impact of crises on the crime situation in Germany, and of course there is a lot of discussion within the group.
How do Criminal Offenses Change in »Crisis« Situations?
© Fraunhofer ITWM
Under the guidance of supervisors from the TU Kaiserslautern and the Fraunhofer ITWM, the groups fiddle their way through the tasks.

Artificial Intelligence in the Context of Reinforcement Learning

Artificial intelligence is currently on everyone's lips and is increasingly being applied to decision-making problems. But what is really hidden behind the term artificial intelligence? This is the question the team is investigating.

The project is dedicated to the field of reinforcement learning, which forms the basis for an area of artificial intelligence. Here, there is an agent that is supposed to find its way around an environment. It explores its environment and is supposed to learn independently from good as well as bad events that it causes by its actions.

 

At the end of the Learning Process, the Team Obtains an AI

It is coping well in her environment and is able to make independent decisions within the framework the team has established.. A good example is Super Mario Bros. for the NES. Here, Mario is the agent who is in a 2D world. If you tell the agent that running to the right is good and dying is bad, the it can learn to complete entire levels.

This learning behavior is inspired by humans and animals, making it seem very natural. In this project, the team will learn the core idea of these learning algorithms and implement a simple artificial intelligence with assistance.

The group is supervised by our colleague Tobias Joosten from the division »Optimization« at Fraunhofer ITWM.

Group Artificial Intelligence in the Context of Reinforcement Learning
© Fraunhofer ITWM
The program focuses on working together as a group every day.
Group Artificial Intelligence in the context of reinforcement learning.
© Fraunhofer ITWM
Highest concentration in the group: What decision-making process does the AI go through?
Group Artificial Intelligence in the context of reinforcement learning.
© Fraunhofer ITWM
This project introduces the core idea of learning algorithms and implements a simple Artificial Intelligence.
 

Interviews

We interviewed our participants about their experiences. Students from each of the three groups talk about their project work, the leisure program and describe their impressions: [Only available in German]

Impressionen

 Gruppenbild Math-Talent-School
© Fraunhofer ITWM
Gruppenbild Math-Talent-School
Group Tour through Fraunhofer ITWM
© Fraunhofer ITWM
Our colleague Ilka Blauth leads the participants through our institute.
Rocket Group: Schoolgirl works cross-legged on formulas
© Fraunhofer ITWM
If you want to land on the moon, you need a lot of space for the right formulas: On the way to the rocket launch, the participants work intently on their project both at their desks and on the ground.
AI exhibition: front page of the comic essay »AI, we need to talk«.
The comic essay »KI, wir müssen reden« (»AI, we need to talk«) at the AI exhibition enlightens participants about the background and diversity of artificial intelligence.
Schoolgirls try out an application at the AI exhibition
© Fraunhofer ITWM
Three schoolgirls try out an application at the AI exhibition at our institute: The AI learns to recognize and assign the numbers they have drawn in.
AI exhibition: traffic example is tested by participants
The various decision-making dilemmas that arise for artificial intelligences in road traffic accidents are exemplified by an interactive application.