Making machine learning transparent, reliable & practical.
My research focuses on explainable artificial intelligence and efficient
computational methods — building models whose decisions can be understood,
trusted, and actually run on real-world hardware.
Let's build things that make sense.
Vita
& education
Current Position
University Researcher · Akademische Rätin
Friedrich-Alexander-Universität (FAU) Erlangen-Nürnberg — Faculty of Engineering. Research on explainable AI and computation coding.
Education
2024 – ongoing
PhD in Engineering (Dr.-Ing.)
Faculty of Engineering, FAU Erlangen-Nürnberg Research focus: artificial intelligence & computation coding
2022 – 2024
Elite Master of Science — Advanced Signal Processing & Communications Engineering
Faculty of Engineering, FAU Erlangen-Nürnberg · Final grade 1.0 Thesis: Linear Computation Coding for Transformer-Based Large Language Models (1.0)
2019 – 2022
Bachelor of Science — Electrical Engineering, Electronics & Information Technology
Faculty of Engineering, FAU Erlangen-Nürnberg · Final grade 1.2 Thesis: Algorithms for Matrix Decomposition for Computational Coding (1.0)
2011 – 2019
Abitur (high school diploma)
Dominicus-von-Linprun-Gymnasium Viechtach · Final grade 1.0
International Experience
2022
Erasmus Semester — Norwegian University of Science & Technology
2016 – 2017
Exchange Year — Camas High School, Washington State, USA
Research
& interests
Working at the intersection of machine learning and hardware-efficient computation — compressing neural networks without losing the ability to explain what they do.
01
Explainable AI
Making the decisions of machine-learning models transparent and trustworthy.
02
Linear Computation Coding
Efficient algorithms for approximating the matrix products at the heart of deep models.
03
Efficient Neural Networks
Compressing networks to run on reconfigurable hardware with minimal cost.
04
Signal Processing & Information Theory
The theoretical backbone connecting communication, coding, and computation.
Selected Publications
2026
J. S. Fröhlich, B. Heinlein, J. U. Claar, H. Rosenberger, V. Belagiannis, and R. R. Müller, “The Confusion is Real: GRAPHIC — A Network Science Approach to Confusion Matrices in Deep Learning,” Transactions on Machine Learning Research (TMLR), 2026.
2025
J. S. Fröhlich, H. Rosenberger, and R. R. Müller, “Spicing up LLMs: the role of PAPRICA pruning in linear computation coding,” in Proc. 2025 IEEE Conference on Artificial Intelligence (CAI), 2025, pp. 360–365.
2025
H. Rosenberger, R. Fischer, J. S. Fröhlich, A. Bereyhi, and R. R. Müller, “Coding for Computation: efficient compression of neural networks for reconfigurable hardware,” 2025 IEEE Statistical Signal Processing Workshop (SSP), 2025.
2023
H. Rosenberger, J. S. Fröhlich, A. Bereyhi, and R. R. Müller, “Linear computation coding: exponential search and reduced-state algorithms,” in Proc. 2023 Data Compression Conference (DCC), 2023, pp. 298–307.
On stage
Speaking & conferences
From workshops to international venues, Johanna regularly presents her
research on explainable, efficient AI — turning dense results into ideas a
room can follow.
NHR Conference '25 High Performance Computing · Göttingen
Teaching
& mentoring
Courses taught and supported at the Faculty of Engineering, FAU Erlangen-Nürnberg.
Information Theory and Coding
Graduate
Fundamentals of Electrical Engineering
Undergraduate
Mathematics
Undergraduate
150
Teaching Hours
500
Students
15
Talks
Awards
& scholarships
Scholarships
Studienstiftung des deutschen Volkes e.V.
Max Weber Program
ARIADNE Technat Mentoring
Dr. Johannes Heidenhain Foundation
Awards
Best Paper Award — IEEE Conference on Artificial Intelligence, 2025
2nd Place — Green ICT Award 2024 (bachelor's thesis)
Semikron Danfoss Woman Award 2023 & Semikron Student Award 2021
Mathematics Abitur Prize — German Mathematical Society
2nd Place — Dr. Hans Riegel Awards (seminar paper “Theremin”)
Off the clock
From algorithms to the dance floor
When she's not compressing neural networks, Johanna is on the parquet. A
passionate ballroom dancer, she now also teaches lessons at a local dance
school — the same love of structure, rhythm, and precision, in a different
key.
Let's collaborate
Open to research & conversations
Working on explainable, efficient AI — and always glad to talk shop. Reach
out and let's find the questions worth answering.
