Matthias Probst

Matthias Probst is a Researcher at the Department of Information Security, Technical University of Munich (TUM), specializing in hardware security and neural network implementations. His work bridges cryptographic engineering with emerging computing paradigms, focusing on vulnerabilities in neuromorphic hardware and side-channel analysis of neural networks.

Research Focus: Probst investigates Side-Channel Analysis techniques against neuromorphic systems, particularly Spiking Neural Networks (SNNs) and Physical Unclonable Functions (PUFs). His recent publications reveal expertise in

• Hardware security countermeasures (DOMREP series)
• Fault injection methodologies (Switch-Glitch)
• Secure neural network acceleration
• Quantum-resistant hardware implementations

His work demonstrates how timing and electromagnetic emanations can extract secrets from emerging hardware architectures.

Publication Trends: Analyzing Probst's 2019-2025 publications shows increasing focus on neural network security (60% of recent work), with growing emphasis on neuromorphic hardware (40% since 2022). His research uniquely combines traditional side-channel analysis with next-generation computing architectures, revealing critical vulnerabilities in spiking neuron implementations and PUF-based security primitives. Key themes include timing-based attacks on loop PUFs, electromagnetic fault injection sweet spots, and masked neural network accelerators.

Collaboration Network: Probst consistently collaborates with TUM's Chair of Information Security team led by Prof. Georg Sigl, particularly with Manuel Brosch and Michael Gruber. His 2024 EMDRIVE Architecture paper involves cross-institutional work spanning embedded diagnostics to edge computing.