Logo
Graphs in Artificial Intelligence
and
Neural Networks

GAIN Group Members

Currently, the GAIN-group consists of group leader, Ph.D. students,,,, research associate and student assistant Anthony Orellana.

In addition,Prof. Dr. Bernhard Sick, head of the Intelligent Embedded Systems department of the University of Kassel, is our mentor. And, senior scientist in Machine Learning and group leader at Fraunhofer IEE, is project leader of GAIN at our project partner Fraunhofer IEE.

Dr. rer. nat. Josephine Thomas

Josephine is the leader of the junior research group GAIN.

She is a trained physicist and gained her doctorate with a thesis on non-linear dimension reduction applied to a topic of complex network theory.

She choose to work on Machine Learning and specifically on GNNs because she likes to do basic reseach but also see it go to a useful application quickly and because graphs are just cool: they can describe the formation of the universe as well as the internet, what more can a scientist want?

Her focus in GAIN next to supervising PhD students is the explainability of GNN algorithms.



Areas of Interest:

Graphs, deep learning, network geometry, quantum physics, renewable energies.



E-mail:

Phone: +49 561 804 6061

PGP-key: Download here

M.Sc. Silvia Beddar-Wiesing

Silvia is working towards her Ph.D. in the GAIN group with Dr. Josephine Thomas and Prof. Dr. Bernhard Sick as her supervisors.

She graduated with a B.Sc. in Mathematics and a M.Sc. in Computer Science with a specialization in Computational Intelligence and Data Analytics at the University of Kassel.

Her research topic addresses the embedding of structural-dynamic graphs using Neural Networks. The focus is on embeddings that transform graphs with changing nodes and edges over time, i.e., including additions and deletions of nodes and edges. In particular, the goal is to design a model based on Temporal Point Processes that represents structural-dynamics in graphs as density function over time and, thus, provides an event prediction procedure. Furthermore, Silvia is interested in comparing dynamic graphs and analyzing the expressivity of GNNs using the Weisfeiler-Lehman test.



Areas of Interest:

Pattern Recognition in Temporal and Structural Data, Expressivity of Neural Networks, Similarity Measures, Data/Information/Knowledge Fusion.



E-mail:

Phone: +49 561 804 6454

PGP-key: Download here

Follow: in

M.Sc. Alice Moallemy-Oureh

I am Alice and I am currently in the third year of my doctorate in the GAIN Group. Prior, I graduated with a M.Sc. in Mathematics and its applications to Computer Science from the university of Kassel. My specialisation in Mathematics lies in Algorithmic Commutative Computer Algebra and Geometry. In Computer Science, I focused in Logic and Data Analytics.

I am particularly interested in developing Graph Neural Networks, which are designed to master the handling of attribute dynamic graphs in continuous time representation. The first thing I would like to focus on is the so-called node embedding of such graphs, as this can be used to keep almost unlimited application possibilities open. The plan is to create more efficient wind power forecasts with the help of such a model.

In general, I am a very open, honest, and enthusiastic person who likes to meet new people and is up for any discussion - so don't be afraid to chat me up.



Areas of Interest:

Geometric Deep Learning, Machine Learning, Pattern Recognition, Commutative Computeralgebra, Geometry, IT-security.



E-mail:

Phone: +49 561 804 6363

PGP-key: Download here

Follow: in

M.Sc. Clara Holzhüter

I have joined the GAIN group in 2022 after I graduated with B.Sc. and M.Sc. in Computer Science at the University of Göttingen. During my studies, I specialized in Computational Neuroscience, especially in Deep Learning and Computer Vision.

The research area of my Ph.D. in the GAIN group is the application of GNNs in power grid applications. Due to the inherent graph structure of electricity networks, GNNs are promising methods with regard to representation learning and information extraction for different use cases. My goal is to develop GNNs which enhance the efficiency and safety of electric networks by directly making use of their topology. For this purpose, I aim to combine Reinforcement Learning with GNNs to learn to control the power grid with regard to its topology. The field of graphs which change their topology or attributes over time is of particular interest for me, since power grids are highly dynamic systems.

As a Ph.D. student at the Fraunhofer IEE my research also addresses the application of such methods to real world problems, particularly in the field of energy applications.



Areas of Interest:

Graph Neural Networks, Deep Learning, Reinforcement Learning, Electricity Grids, Machine Learning, Graph Theory, Renewable Energies



E-mail (Uni):

PGP-key (Uni): Download here

E-mail (Fraunhofer):

Follow: in

M.Sc. Pascal Plettenberg

I have joined the GAIN group in 2024. I graduated with an M. Sc. in Physics from the University of Kassel and during my studies, I specialized on Machine Learning applications in Theoretical Physics and Materials Science.

Currently, I am aiming for my Ph.D. and my research topic is the application of GNNs to the design and optimization of Printed Circuit Boards (PCBs). Electronic circuits can be naturally represented as graphs and therefore, GNNs are a promising method for optimizing them in order to obtain electronic devices that are more efficient and reliable. For this purpose, my goal is to combine GNNs with other approaches, such as Human-in-the-Loop and Physics-informed Machine Learning.



Areas of Interest:

Deep Learning, Graph Neural Networks, Computer Vision, Electronics, Materials Science, Quantum Physics, Fuel Cells



E-mail:

PGP-key: Download here

Follow: in

Cooperations

We are funded by the Federal Ministry of Education and Research Germany (BMBF) under the following funding codes:
01IS20047A, according to the 'Policy for the funding of female junior researchers in Artificial Intelligence'.
020E-100626677, within the '7. Energieforschungsprogramm'.
16ME0877, according to the KMU-innovativ' guideline.

The responsibility for the content of this website or of any publication lies with the author.