Master Thesis in Error Correction for Binary Neural Networks in Noisy Hardware and Space Environment

The DLR Institute of Communications and Navigation is dedicated to mission-oriented research in selected areas of communications and navigation. Its work ranges from the theoretical foundations to the demonstration of new procedures and systems in a real environment and is embedded in DLR's Space, Aeronautics, Transport, Security and Digitalization programmes.

What to expect

The Information Transmission Group was established within the Satellite Networks Department of the Institute of Communications and Navigation to investigate techniques for reliable and secure transmission, processing and storage of information. Grounded on a firm theoretical background, we target the development of algorithms for satellite and (more generally) for wireless communication systems, with emphasis on forward error correction schemes, physical layer techniques and (massive) multiple access. Specific application areas of interest are currently satellite/cellular IoT systems, high-throughput wireless links (including free-space optical communications), as well as high-mobility channels and reliable processing and storage of information.

Your tasks

• investigation of coding-theoretic methods to increase the robustness of binary neural networks implemented on noisy hardware platforms, such as memristor crossbars or radiation-prone memories used in space applications
• research involves the analytical modeling of the noisy hardware environment
• implementation and evaluation of various decoding strategies (e.g., Belief Propagation, Approximate Message Passing, and hard-decision decoders)
• exploring the trade-offs between decoding complexity and robustness

Your profile

• Solid background in digital communications, information, and coding theory.
• Fundamental understanding of neural networks.
• Advance programming skills.
• Excellent academic records.

We look forward to getting to know you!

If you have any questions about this position (Vacancy-ID 5366) please contact:

Dr. Federico Clazzer
Tel.: +49 (0) 8153 281120