TUM Logo

Florian Kohlmayer

Dipl.-Inf. Florian Kohlmayer


now at TUM School of Medicine/Informatics, Ph.D. Student)

Research Interests

In translational medical research, complex data from heterogeneous sources have to be integrated, e.g. clinical data, study data, bio-bank data, images and the results of "omics" analyses (genomics, proteomics, metabolomics, etc.). IT security and data safety poses significant challenges as data are highly sensible and their use may be restricted to specific purposes only. Patient privacy has to be guaranteed, and access privileges need to be kept to a minimum. Informed consent and anonymity are principles of central relevance.

The goal of my dissertation will be to perform research on security and pseudonymity concepts for health information systems that ensure a maximum degree of integrity on one hand and anonymity on the other hand. Aspects like k-anonymity and prevention of linkage need to be explored.

Existing data protection concepts have to be evaluated and further developed regarding requirements for integrating various types of data sources. Single-Sign-on, identity and access rights as well as information flow management will play an important role from the application perspective. On the research side, appropriate security models and security primitives need to be investigated to allow one to rigorously argue about the security of the developed concepts and architectures against various attacks and forms of abuse.

Current project: ARX - Powerful Data Anonymization - K-Anonymity, L-Diversity, T-Closeness Implementation in Java
"This project aims at providing a comprehensive, open source anonymization framework for sensitive personal data. It is able to alter the data in a way that guarantees minimal information loss while making sure that the transformed data adheres to well-known privacy criteria, such as k-anonymity, l-diversity or t-closeness (coming soon: d-presence).

My dissertation is covered under the umbrella of the Graduate School of Information Science in Health. Further information can be found at: http://gsish.tum.edu/

(private website)


2012 Flash: Efficient, Stable and Optimal K-Anonymity
Highly Efficient Optimal K-Anonymity For Biomedical Datasets
2010 Konzept für ein deutschlandweites Krankheitsnetz am Beispiel von mitoREGISTER