Marcel Gehrke

Marcel Gehrke

Auto-generated profile, based on publicly available information

Marcel Gehrke

University of Hamburg

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 May 2018
Verified
Book Chapter
Closed access
PDF file
MGMarcel Gehrke
Marcel Gehrke
TBTanya Braun
Tanya Braun
RMRalf Möller
Ralf Möller

21 February 2019
Verified
Book Chapter
Closed access
PDF file
MGMarcel Gehrke
Marcel Gehrke
TBTanya Braun
Tanya Braun
RMRalf Möller
Ralf Möller
AWAlexander Waschkau
Alexander Waschkau
CSChristoph Strumann
Christoph Strumann
JSJost Steinhäuser
Jost Steinhäuser

2 December 2024
Verified
Conference Paper
Open access
JLJohannes Liebenow
Johannes Liebenow
YSYara Schütt
Yara Schütt
TBTanya Braun
Tanya Braun
MGMarcel Gehrke
Marcel Gehrke
FTFlorian Thaeter
Florian Thaeter
EMEsfandiar Mohammadi
Esfandiar Mohammadi
Clustering is an important tool for data exploration where the goal is to subdivide a data set into disjoint clusters that fit well into the underlying data structure. When dealing with sensitive data, privacy-preserving algorithms aim to approximate the non-private baseline while minimising the leakage of sensitive information. State-of-the-art privacy-preserving clustering algorithms tend to output clusters that are good in terms of the standard metrics, inertia, silhouette score, and clustering accuracy, however, the clustering result strongly deviates from the non-private KMeans baseline. In this work, we present a privacy-preserving clustering algorithm called DPM that recursively separates a data set into clusters based on a geometrical clustering approach. In addition, DPM estimates most of the data-dependent hyper-parameters in a privacy-preserving way. We prove that DPM preserves Differential Privacy and analyse the utility guarantees of DPM. Finally, we conduct an extensive empirical evaluation for synthetic and real-life data sets. We show that DPM achieves state-of-the-art utility on the standard clustering metrics and yields a clustering result much closer to that of the popular non-private KMeans algorithm without requiring the number of classes.

24 April 2019
Verified
Book Chapter
Closed access
PDF file
MGMarcel Gehrke
Marcel Gehrke
TBTanya Braun
Tanya Braun
RMRalf Möller
Ralf Möller

10 November 2023 in KI - Künstliche Intelligenz
Verified
Journal Article
Open access
PDF file
MGMarcel Gehrke
Marcel Gehrke
Processes in our world are of a temporal probabilistic relational nature. An epidemic is an example of such a process. This dissertation abstract uses the scenario of an epidemic to illustrate the lifted dynamic junction tree algorithm (LDJT), which is a temporal probabilistic relational inference algorithm. More specifically, we argue that existing propositional temporal probabilistic inference algorithms are not suited to model an epidemic, i.e., without accounting for the relational part, and present how LDJT uses the relational aspect. Additionally, we illustrate how LDJT preserves groups of indistinguishable objects over time and have a look at LDJT from a theoretical side.

13 June 2024 in KI - Künstliche Intelligenz
Verified
Journal Article
Open access
PDF file
MGMarcel Gehrke
Marcel Gehrke
JLJohannes Liebenow
Johannes Liebenow
EMEsfandiar Mohammadi
Esfandiar Mohammadi
TBTanya Braun
Tanya Braun
Privacy-preserving inference aims to avoid revealing identifying information about individuals during inference. Lifted probabilistic inference works with groups of indistinguishable individuals, which has the potential to prevent tracing back a query result to a particular individual in a group. Therefore, we investigate how lifting, by providing anonymity, can help preserve privacy in probabilistic inference. Specifically, we show correspondences between k-anonymity and lifting and present s-symmetry as an analogue as well as PAULI, a privacy-preserving inference algorithm that ensures s-symmetry during query answering.