TY  - GEN
A1  - Ziegler, Joceline
A1  - Pfitzner, Bjarne
A1  - Schulz, Heinrich
A1  - Saalbach, Axel
A1  - Arnrich, Bert
T1  - Defending against Reconstruction Attacks through Differentially Private Federated Learning for Classification of Heterogeneous Chest X-ray Data
T2  - Zweitveröffentlichungen der Universität Potsdam : Reihe der Digital Engineering Fakultät
N2  - Privacy regulations and the physical distribution of heterogeneous data are often primary concerns for the development of deep learning models in a medical context. This paper evaluates the feasibility of differentially private federated learning for chest X-ray classification as a defense against data privacy attacks. To the best of our knowledge, we are the first to directly compare the impact of differentially private training on two different neural network architectures, DenseNet121 and ResNet50. Extending the federated learning environments previously analyzed in terms of privacy, we simulated a heterogeneous and imbalanced federated setting by distributing images from the public CheXpert and Mendeley chest X-ray datasets unevenly among 36 clients. Both non-private baseline models achieved an area under the receiver operating characteristic curve (AUC) of 0.940.94 on the binary classification task of detecting the presence of a medical finding. We demonstrate that both model architectures are vulnerable to privacy violation by applying image reconstruction attacks to local model updates from individual clients. The attack was particularly successful during later training stages. To mitigate the risk of a privacy breach, we integrated Rényi differential privacy with a Gaussian noise mechanism into local model training. We evaluate model performance and attack vulnerability for privacy budgets ε∈{1,3,6,10}�∈{1,3,6,10}. The DenseNet121 achieved the best utility-privacy trade-off with an AUC of 0.940.94 for ε=6�=6. Model performance deteriorated slightly for individual clients compared to the non-private baseline. The ResNet50 only reached an AUC of 0.760.76 in the same privacy setting. Its performance was inferior to that of the DenseNet121 for all considered privacy constraints, suggesting that the DenseNet121 architecture is more robust to differentially private training.
T3  - Zweitveröffentlichungen der Universität Potsdam : Reihe der Digital Engineering Fakultät - 14 
KW  - federated learning
KW  - privacy and security
KW  - privacy attack
KW  - X-ray
Y1  - 2023
U6  - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus4-581322
IS  - 14
ER  - 
TY  - GEN
A1  - Lever, Fabiano
A1  - Mayer, Dennis
A1  - Metje, Jan
A1  - Alisauskas, Skirmantas
A1  - Calegari, Francesca
A1  - Düsterer, Stefan
A1  - Feifel, Raimund
A1  - Niebuhr, Mario
A1  - Manschwetus, Bastian
A1  - Kuhlmann, Marion
A1  - Mazza, Tommaso
A1  - Robinson, Matthew Scott
A1  - Squibb, Richard J.
A1  - Trabattoni, Andrea
A1  - Wallner, Måns
A1  - Wolf, Thomas J. A.
A1  - Gühr, Markus
T1  - Core-level spectroscopy of 2-thiouracil at the sulfur L1 and L2,3 edges utilizing a SASE free-electron-laser
T2  - Postprints der Universität Potsdam : Mathematisch-Naturwissenschaftliche Reihe
N2  - In this paper, we report X-ray absorption and core-level electron spectra of the nucleobase derivative 2-thiouracil at the sulfur L1- and L2,3-edges. We used soft X-rays from the free-electron laser FLASH2 for the excitation of isolated molecules and dispersed the outgoing electrons with a magnetic bottle spectrometer. We identified photoelectrons from the 2p core orbital, accompanied by an electron correlation satellite, as well as resonant and non-resonant Coster–Kronig and Auger–Meitner emission at the L1- and L2,3-edges, respectively. We used the electron yield to construct X-ray absorption spectra at the two edges. The experimental data obtained are put in the context of the literature currently available on sulfur core-level and 2-thiouracil spectroscopy.
T3  - Zweitveröffentlichungen der Universität Potsdam : Mathematisch-Naturwissenschaftliche Reihe - 1180 
KW  - X-ray
KW  - photoelectron
KW  - sulfur
KW  - thiouracil
KW  - nucleobases
KW  - Coster–Kronig
KW  - Auger–Meitner
KW  - NEXAFS
KW  - FLASH
Y1  - 2021
U6  - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus4-524091
SN  - 1866-8372
IS  - 21
ER  -