KCDS Virtual Open House 2025 - Spring Edition

The KCDS Virtual Open House is currently being updated.

Right now, you can already find out about

the doctoral projects in this application round
- Project 01: How to achieve “entity awareness” in deep learning models for hydrological applications? (DAAD scholarship)
- Project 02: Non-intrusive calculation of homogenized material properties for non-periodic and non-linear microstructures (DAAD scholarship)
- Project 03: Doctoral Project in the group of André Platzer, the Alexander von Humboldt Professor for Logic of Autonomous Dynamical Systems (doctoral position, employment at KIT)
- Project 04: Replacement of chemistry simulation by machine learning in an Earth system model (doctoral position, employment at KIT)
requirements for applicants

Project 01: How to achieve “entity awareness” in deep learning models for hydrological applications? (DAAD scholarship)

A symbolic image of entity awareness in deep learning models for hydrological applications, generated by an AI tool

Project Description

Accurate prediction of hydrological processes is crucial for effective water resource management. Recent advancements in machine learning, particularly deep learning (DL), have shown great potential for modeling various fluxes and state variables in hydrological systems. One of the most successful architectures is the Long Short-Term Memory (LSTM) network. While LSTMs often outperform traditional process-based hydrological models, their statistical properties have received limited attention. Moreover, there is evidence that LSTMs, as applied in hydrology, struggle to effectively learn from time-invariant, static information making out-of-sample predictions uncertain and challenging.

This project aims to improve the generalization of current DL models in hydrology to unseen sites. To achieve this, we propose leveraging a large dataset from process-based hydrological modeling to explore how embedding system characteristics within DL architectures can enhance the representation of structural attributes. By adopting a functional regression perspective and a mathematical analysis of the statistical properties of the equipped neural network architectures, we aim to improve the models’ understanding of how time-invariant and slowly varying features influence hydrological responses. Our overarching goal is to develop robust and generalizable DL models for Earth system sciences, with a particular focus on improving performance in unseen or changing environments.

The work program includes the development of an exemplary toy model using the process-based hydrological model as a testbed and the development of a new DL model based on mathematical principles, which leverages a functional regression perspective with the dynamic LSTM approach.

Special requirements for applicants:

Solid mathematical background, especially in probability theory and statistics
Either hydrological background and high motivation to implement and apply DL methods or good DL background and curiosity for hydrological applications.

Requirements for applicants

Project 01 is a scholarship funded by the German Academic Exchange Service (DAAD) in the DAAD Graduate School Scholarship Program.

Special requirements for applicants apply:

Completed Master’s degree (or equivalent) at the starting date of the scholarship / preparatory German course
Graduation no more than six years prior to nomination (exceptions apply, see DAAD website)
No residency in Germany for more than the past 15 months before nomination ("mobility rule")
No completed PhD degree
Proficiency in English (German is optional)

Funding

Funding for this scholarship is provided by the German Academic Exchange Service (DAAD).

Duration of the funding: Up to 4 years
Monthly payments of 1,300 euros (note: this will be supplemented by KIT with a part-time job with an additional income of EUR 556 gross per month)
Payments towards health, accident and personal liability insurance cover
Travel lump sum (depending on country of origin)
Annual research allowance of 460 euros

Under certain circumstances, grant holders may receive the following additional benefits:

monthly rent subsidy
monthly allowance for accompanying members of family (about 200 euros child allowance per child; about 275 euros marriage allowance)

KCDS Supervisors for Project 01
Title Name	Role at KCDS	Homepage
Dr. rer. nat. Loritz, Ralf	SEE Supervisor
Prof. Dr. Trabs, Mathias	MATH Supervisor, member of the Steering committee	www.math.kit.edu/stoch/~trabs/

Project 02: Non-intrusive calculation of homogenized material properties for non-periodic and non-linear microstructures (DAAD scholarship)

Automotive composite underfloor with macro-scopic features (left) and generic heterogenous micro-structure (right)

Project Description

Fiber-reinforced materials are important in industrial areas with high lightweight design requirements. However, the very heterogeneous structure of such materials makes their simulation cumbersome: Fully resolving microstructures is computationally expensive or even impossible, which is why homogenization approaches are used. These approaches replace heterogeneous materials by piecewise homogeneous ones. In applied mathematics, a different approach to homogenization has become popular in recent years: Instead of computing homogenized material properties, the basis functions of a finite element method are suitably adjusted to better fit to the microstructure. The advantage of such methods is that they work under minimal assumptions on the microstructure. Their drawback is that they are intrusive: They change the basis functions, which makes the implementation of such strategies into available legacy code very complicated.

In this project, we will start from mathematically justified intrusive strategies and adjust them suitably to obtain a non-intrusive method that shall then be implemented into available code. Moreover, we will also cover non-linear material properties, where appropriate linearization strategies are required. Finally, we will compare simulation results with practical experiments to assess the quality of the novel strategy. Here, the Institute for Applied Materials (IAM) will provide additional support.

As a candidate, you should have:

a solid background in applied mathematics
knowledge of the finite element method and its theoretical investigation
basic programming skills (in any programming language)
interest in working in an interdisciplinary environment

Requirements for applicants

Project 02 is a scholarship funded by the German Academic Exchange Service (DAAD) in the DAAD Graduate School Scholarship Program.

Special requirements for applicants apply:

Completed Master’s degree (or equivalent) at the starting date of the scholarship / preparatory German course
Graduation no more than six years prior to nomination (exceptions apply, see DAAD website)
No residency in Germany for more than the past 15 months before nomination ("mobility rule")
No completed PhD degree
Proficiency in English (German is optional)

Funding

Funding for this scholarship is provided by the German Academic Exchange Service (DAAD).

Duration of the funding: Up to 4 years
Monthly payments of 1,300 euros (note: this will be supplemented by KIT with a part-time job with an additional income of EUR 556 gross per month)
Payments towards health, accident and personal liability insurance cover
Travel lump sum (depending on country of origin)
Annual research allowance of 460 euros

Under certain circumstances, grant holders may receive the following additional benefits:

monthly rent subsidy
monthly allowance for accompanying members of family (about 200 euros child allowance per child; about 275 euros marriage allowance)

KCDS Supervisors for Project 02
Title Name	Role at KCDS	Homepage
Prof. Dr.-Ing. Kärger, Luise	SEE Supervisor
TT-Prof. Dr. Maier, Roland	MATH Supervisor	www.math.kit.edu/ianm_jrg_numpde/~roland.maier/en

By clicking on the button below, data is loaded from the following external servers:

www.youtube.com/embed/aF_n1HvIr-A

Personal data (e.g. your IP address) is transmitted to these external servers.

The web page would like to load the following external sources:

www.youtube.com/embed/aF_n1HvIr-A

These are not loaded for data protection reasons! Activate JavaScript if you want to load this content. You only see the local content.

Project 03: Doctoral Project in the group of André Platzer, the Alexander von Humboldt Professor for Logic of Autonomous Dynamical Systems (doctoral position, employment at KIT)

Project description

The group of André Platzer, the Alexander von Humboldt Professor for Logic of Autonomous Dynamical Systems, in the Computer Science Department at KIT, Karlsruhe, is recruiting a PhD Student / Doctoral Researcher (full-time, about €4200-€4800 gross by TVL E13 depending on experience). Exceptionally qualified applicants for Postdoc positions may be considered as well.

Our research group develops the logical foundations for cyber-physical systems and practical theorem proving tools such as KeYmaera X for analyzing and correctly building such systems. Our techniques are used to analyze the safety of autonomous cars, airplanes and collision avoidance systems in aerospace applications, robotics, and train control as well as for provably safe AI. Your exciting research can have a direct impact on making the world a better place!

Successful applicants to the Alexander von Humboldt Professor group bring the following:

Excellent master's degree in computer science, mathematics or related subjects.
Strong demonstrable commitment to research.
Strong background in proof-based sciences such as logic, mathematics, theorem proving, formal methods, or programming language principles.
Strong background in mathematics, physics, or engineering.
Proficiency in English, excellent speaking and writing skills.
Experience in software development projects is a plus.
Exceptionally qualified bachelor students, especially from 4 year programs may apply for a fast-track program.

The successful candidate shows initiative and is able to quickly get into new research areas and will be responsible for actively engaging in novel research questions, publishing and communicating research results, advising junior students, implementation of research results in formal methods or theorem proving tools, demonstrating their applicability in cyber-physical systems applications.

Faculty / Division: Alexander von Humboldt Professor group on Logic of Autonomous Dynamical Systems, KIT Department of Informatics

Group: The Alexander von Humboldt Professorship, Germany's most highly endowed international research award, is endowed with up to five million euros. With this award, the Alexander von Humboldt Foundation honors internationally leading researchers of all disciplines.

Institute: Institute of Information Security and Dependability (KASTEL)
Institute for Reliability of Autonomous Dynamical Systems (VADS)
of the Karlsruhe Institute of Technology (KIT), Germany

Funding

This is a 1 full-time equivalent "E13 TV-L" position under applicable regulations of KIT according to the public sector salary system at German universities and research centers. It means a yearly gross income of more than 57.000 € and results in more than 35.000 € net income after taxes, own contributions to health insurance, pension insurance, unemployment insurance etc.

KCDS Supervisor for Project 03
Title Name	Role at KCDS	Homepage
Prof. Dr. Platzer, André	MATH Supervisor	logic.kastel.kit.edu/andre.html

Project 04: Replacement of chemistry simulation by machine learning in an Earth system model (doctoral position, employment at KIT)

Project description

Earth system modeling plays a central role in climate change research. Earth system models (ESMs) describe processes such as the change in atmospheric dynamics and temperatures and the distribution of aerosols and chemical substances and their interaction with climate change. One of the most computationally intensive processes in ESMs is the calculation of chemistry.
The goal of this doctoral research is to show that the simulation of chemistry within an ESM can be replaced by an AI model. By combining domain knowledge from earth sciences with state-of-the-art ML techniques, the project seeks to develop novel hybrid ESM models that incorporate physical laws and KI driven approaches. Classical climate models use complex partial differential equations (as e.g. Navier-Stokes equations, continuity equations, radiative transfer equations, thermodynamic energy equations, or tracer transport and diffusion equations) to simulate the state of the atmosphere and ocean. More recent ESMs integrate additional processes to the
classical climate models, as for example the simulation of chemistry, which is based on a stiff set of ODEs. We aim to replace these difficult and expensive (in terms of computing time) to solve equations by neural networks that conserve the physical constrains of the system. Thereby we aim to train all chemical tracers of the ESM with the help of an AI model and integrate afterwards a neural network into the ESM.

It will be investigated and evaluated if this enables a realistic simulation and thus improves the performance of the ESM.
The work involves an interdisciplinary approach and includes mathematics, atmospheric science and computer science. Corresponding contact persons are available.

Tasks of the thesis include:

Performance analysis of the Earth System model code at different resolutions on a High Performance Computing (HPC) system
Creation of a concept for replacing the chemistry simulation with a suitable AI model (such as Transformer, LSTM, CNN) including the identification of suitable features dimension reduction, and hyperparameter tuning
Implementation and evaluation of the procedure and investigation of suitable metrics
Evaluation of the AI model integrated into the Earth System Model (including parallelization)

Necessary requirements:

Completed studies (master) in mathematics
Knowledge of current deep learning frameworks (e.g., PyTorch or Tensorflow)
Programming skills (e.g. Fortran, C++, Python)
Ability to work and publish in a targeted and scientific manner.
Good communication and presentation skills and willingness and ability to work a team
Good writing and oral communication skills in English

Optional requirements:

Experience in working with climate/earth system models

Funding

This is a 0.75 full-time equivalent "E13 TV-L" position under applicable regulations of KIT according to the public sector salary system at German universities and research centers. It means a yearly gross income of more than 43.000 € and results in more than 27.000 € net income after taxes, own contributions to health insurance, pension insurance, unemployment insurance etc.

KCDS Supervisors for Project 04
Title Name	Role at KCDS	Homepage
Prof. Dr. Frank, Martin	Scientific Speaker, MATH Supervisor
Dr. Hörter, Jasmin
Dr. Kirner, Oliver	MATH Supervisor	www.scc.kit.edu/personen/oliver.kirner.php
Dr. Meyer, Jörg	SEE Supervisor	www.scc.kit.edu/en/staff/joerg.meyer.php

Requirements for applicants

General requirements

For all applicants, the following requirements apply:

We are looking for excellent graduates holding master degrees, received by the start of their doctoral studies at the latest, in mathematics or the SEE disciplines (natural sciences, engineering, economics), who have sufficient proficiency in mathematics and are interested in joint research that revolves around computational methods such as mathematical models, simulation methods and data science techniques.
As an international research school, we require our doctoral researchers to have good writing and oral communication skills in English (German is optional).

Requirements for DAAD Graduate School Scholarships

Project 1 and 2 are scholarships funded by the German Academic Exchange Service (DAAD) in the DAAD Graduate School Scholarship Program.

Special requirements for applicants apply:

Completed Master’s degree (or equivalent) at the starting date of the scholarship / preparatory German course
Graduation no more than six years prior to nomination (exceptions apply, see DAAD website)
No residency in Germany for more than the past 15 months before nomination ("mobility rule")
No completed PhD degree
Proficiency in English (German is optional)

Requirements for doctoral positions (employment at KIT)

Project 3 and 4 are doctoral positions (according to the public sector salary system at German universities and research centers), with the type of position specified in the project description.

The following requirements apply:

Completed Master’s degree (or equivalent) at the starting date of doctoral position
No completed PhD degree
Proficiency in English (German is optional)
These positions are also open to candidates with residency in Germany (no "mobility rule"). For these positions, we welcome applications of both international candidates and candidates with residency in Germany!

KCDS Virtual Open House 2025 - Spring Edition

Project 01: How to achieve “entity awareness” in deep learning models for hydrological applications? (DAAD scholarship)

Project Description

Requirements for applicants

Funding

Project Video

Project 02: Non-intrusive calculation of homogenized material properties for non-periodic and non-linear microstructures (DAAD scholarship)

Project Description

Requirements for applicants

Funding

Project 03: Doctoral Project in the group of André Platzer, the Alexander von Humboldt Professor for Logic of Autonomous Dynamical Systems (doctoral position, employment at KIT)

Project description

Funding

Project 04: Replacement of chemistry simulation by machine learning in an Earth system model (doctoral position, employment at KIT)

Project description

Funding

Requirements for applicants

General requirements

Requirements for DAAD Graduate School Scholarships

Requirements for doctoral positions (employment at KIT)