Revolution in information transmission: TU Dresden receives millions in funding!

Revolution in information transmission: TU Dresden receives millions in funding!

The Technical University of Dresden (TUD) has achieved significant success in the field of research. Prof. Jürgen Czarske, head of the professorship for measurement and sensor system technology, has received a funding commitment of around 1.5 million euros as part of the Reinhart Koselleck program of the German Research Foundation (DFG). This project, titled “Physics-Informed Deep Learning Systems for Secure Information Transmission with Multimode Fibers” (Phys-Deep-Fiber), aims to improve fiber optic information transmission using physics-based neural networks. This second Koselleck funding for Czarske since 2014 represents another milestone in the development of technologies in the field of energy and process engineering, in which he has extensive experience.

The project addresses various highly topical topics such as the explainability of AI applications, the energy requirements of AI solutions and the possible applications in optical communication and sensor technology. In particular, the interfaces to areas such as automated driving, intelligent microscopy for biomedicine and quantum communication are being examined as part of the research work. The importance of fiber optic communication for the Internet and the focus on data security and energy-saving technologies underline the relevance of the project in an increasingly digitalized world.

Innovative approaches in AI research

A central element of the project is the development of XAI-based measurement systems that combine data-driven algorithms with physical models. This allows optical neural networks to be trained without the use of power-hungry GPUs, which is both ecologically sensible and economically advantageous. This innovation represents an advance in the energy efficient use of technologies that are becoming increasingly important in today's research.

In the interdisciplinary discussion about neural networks and their foundations, the DFG points out that there are already internationally visible research groups in Germany, including a new priority program called “Theoretical Foundations of Deep Learning”, which promotes interdisciplinary collaborations. One of the main reasons for this focus is the lack of theoretical foundations for neural networks, despite their wide application in the areas of machine learning and generative artificial intelligence, such as systems like ChatGPT, which are increasingly becoming part of our everyday lives.

Successfully combining research and education

In parallel to the current research projects, the importance of solid training in the field of neural networks is emphasized. The book “Neural Networks”, which will be published in its 3rd edition in 2025, offers beginners a well-founded introduction to the basics of the technology behind deep learning and machine learning. Topics such as learning algorithms, programming in Python and ethical issues are covered in detail and are an important part of the learning success for future professionals in this exciting field.

Through the combination of sustainable research, innovative technology and in-depth training, the Technical University of Dresden is sending a clear signal for the future of science in the field of artificial intelligence and energy-efficient systems. A recipient of numerous awards including the Joseph Fraunhofer Award and the Laser Instrumentation Award, Prof. Czarske leads efforts to develop high-quality and trustworthy AI applications capable of addressing the complex challenges of the modern world.

For further information please contact Prof. Jürgen Czarske: Telephone: +49 351 463-34803.