KKKatharina Kohls

Katharina Kohls

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Katharina Kohls has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    AI-Aware Physical-Layer Security for Trustworthy 6G: Threats, Evidence, and Assurance Gaps

    by Jan Herbst, Wenqing Dai, Marten Fischer, Stefan Roth, Jan Petershans, Robin Müller, Stefan Köpsell, Ghazal Bagheri, Christoph Lipps, Kai Jansen, et al.

    Sixth-generation (6G) wireless networks are evolving toward Artificial Intelligence (AI)-native architectures in which learning-based functions enter physical-layer processing, radio access, edge platforms, and network control. These functions can improve adaptation and spectrum use, but training data, model updates, telemetry, and control interfaces also create attack surfaces. In safety-critical environments, failures may cause not only communication disruption but physical harm, operational instability, or regulatory non-compliance. This survey considers Physical Layer Security (PLS) in its broad sense, encompassing confidentiality, authentication, availability and resilience, and sensing or location privacy. Within this scope, we examine three relationships between AI and PLS: AI used to strengthen PLS mechanisms, AI used to undermine their assumptions or defenses, and the security of learning components embedded in PLS control loops. The literature is organized by security objective, attacker capability, lifecycle stage, and experimental evidence. The selected evidence illustrates that mechanism-level improvements alone are insufficient to establish end-to-end security. Exposed models and telemetry, programmable propagation, distribution shift, and closed-loop timing introduce assurance requirements not captured by secrecy rate or classifier accuracy alone. We therefore identify gaps in adversarial over-the-air (OTA) validation, cross-layer threat modeling, reproducible Radio-Frequency (RF) datasets, and lifecycle evidence for body-centric and time-critical cyber-physical systems. Addressing these gaps requires evaluations that connect physical assumptions and learned decisions to operational consequences and recovery behavior.

  • Open Access

    Adversarial Attacks Against Automatic Speech Recognition Systems via Psychoacoustic Hiding

    by Lea Schönherr, Katharina Kohls, Steffen Zeiler, Thorsten Holz, Dorothea Kolossa

    Voice interfaces are becoming accepted widely as input methods for a diverse set of devices. This development is driven by rapid improvements in automatic speech recognition (ASR), which now performs on par with human listening in many tasks. These improvements base on an ongoing evolution of DNNs as the computational core of ASR. However, recent research results show that DNNs are vulnerable to adversarial perturbations, which allow attackers to force the transcription into a malicious output. In this paper, we introduce a new type of adversarial examples based on psychoacoustic hiding. Our attack exploits the characteristics of DNN-based ASR systems, where we extend the original analysis procedure by an additional backpropagation step. We use this backpropagation to learn the degrees of freedom for the adversarial perturbation of the input signal, i.e., we apply a psychoacoustic model and manipulate the acoustic signal below the thresholds of human perception. To further minimize the perceptibility of the perturbations, we use forced alignment to find the best fitting temporal alignment between the original audio sample and the malicious target transcription. These extensions allow us to embed an arbitrary audio input with a malicious voice command that is then transcribed by the ASR system, with the audio signal remaining barely distinguishable from the original signal. In an experimental evaluation, we attack the state-of-the-art speech recognition system Kaldi and determine the best performing parameter and analysis setup for different types of input. Our results show that we are successful in up to 98% of cases with a computational effort of fewer than two minutes for a ten-second audio file. Based on user studies, we found that none of our target transcriptions were audible to human listeners, who still understand the original speech content with unchanged accuracy.