Ephraim Zimmer
YOU?
Author Swipe
“It’s Not My Data Anymore”: Exploring Non-Users’ Privacy Perceptions of Medical Data Donation Apps Open
This paper contributes an in-depth investigation (N=24) of privacy perceptions in the context of medical data donation apps. Medical data donation refers to the act of voluntarily sharing medical data with research institutions, which play…
Anonify: Decentralized Dual-level Anonymity for Medical Data Donation Open
Medical data donation involves voluntarily sharing medical data with research institutions, which is crucial for advancing healthcare research. However, the sensitive nature of medical data poses privacy and security challenges. The primar…
Label-Aware Aggregation for Improved Federated Learning Open
Federated Averaging (FedAvg) is the most common aggregation method used in Federated learning, which performs a weighted averaging of the updates based on the sizes of the individual datasets of each client. A raising discussion in the res…
Mitigating Intersection Attacks in Anonymous Microblogging Open
Anonymous microblogging systems are known to be vulnerable to intersection attacks due to network churn. An adversary that monitors all communications can leverage the churn to learn who is publishing what with increasing confidence over t…
Federated Learning Attacks Revisited: A Critical Discussion of Gaps, Assumptions, and Evaluation Setups Open
Deep learning pervades heavy data-driven disciplines in research and development. The Internet of Things and sensor systems, which enable smart environments and services, are settings where deep learning can provide invaluable utility. How…
User-Level Label Leakage from Gradients in Federated Learning Open
Federated learning enables multiple users to build a joint model by sharing their model updates (gradients), while their raw data remains local on their devices. In contrast to the common belief that this provides privacy benefits, we here…
Federated Learning Attacks Revisited: A Critical Discussion of Gaps, Assumptions, and Evaluation Setups Open
Federated learning (FL) enables a set of entities to collaboratively train a machine learning model without sharing their sensitive data, thus, mitigating some privacy concerns. However, an increasing number of works in the literature prop…
Insiders Dissected: New Foundations and a Systematisation of the Research on Insiders Open
The insider threat is often cited as one of the most challenging threats for security practitioners. Even though this topic is receiving considerable attention, two main problems remain unsolved. First, research on insider threats is focus…
User Label Leakage from Gradients in Federated Learning. Open
Federated learning enables multiple users to build a joint model by sharing their model updates (gradients), while their raw data remains local on their devices. In contrast to the common belief that this provides privacy benefits, we here…
User-Level Label Leakage from Gradients in Federated Learning Open
Federated learning enables multiple users to build a joint model by sharing their model updates (gradients), while their raw data remains local on their devices. In contrast to the common belief that this provides privacy benefits, we here…
PEEPLL Open
Pseudonymisation provides the means to reduce the privacy impact of\nmonitoring, auditing, intrusion detection, and data collection in general on\nindividual subjects. Its application on data records, especially in an\nenvironment with add…
Hashing of personally identifiable information is not sufficient Open
It is common practice of web tracking services to hash personally identifiable information (PII), e. g., e-mail or IP addresses, in order to avoid linkability between collected data sets of web tracking services and the corresponding users…