Biography:Karol Kozak

Short description: Polish- German computer scientist, entrepreneur, academic and author

Karol Kozak is a Polish- German computer scientist, entrepreneur, academic and author. He is a professor, in charge of Biomedical Computer Science Unit at Medical Faculty, Technische Universität Dresden and Leading Cybersecurity Defense Team at State of Saxony , in Dresden, Germany. He has published one book, 6 books/book chapters and 72 papers in scientific journals with high impact factor in the areas of computational methods for industry, cybersecurity and large scale biomedical data. He is member and advisor of Institute of Applied Informatics in Leipzig. He has been involved in the organization of various conferences, most notably the Medical Informatics, Bioinformatics, Cybersecurity-series of symposia on Intelligent Data Management.

Early life and education

Kozak was born in 1976 in Białogard , Polen. He received his MSc degree in computer science in 2001 from Technical Univeristy Radom/ Exchange Program with Technical University Dresden. He received his Dr. rer. nat. degree from Max Planck Society, Germany in cooperation with Silesian University of Technology, POLAND (in discipline: AI, machine learning, biology). Karol Kozak was awarded in 2012 as Junior Professor for Bioinformatics at Philips Marburg University, GERMANY. He became doctor habilitatus in the field of databases for medicine and biotechnology. He became in 2015 a professor title at Medical Faculty, Technical University Dresden in collaboration with Fraunhofer Society.

Career

Kozak started his academic career at as a PhD Student at Max Planck Society in 2002, and as a researcher at the Medical Faculty, TU Dresden in 2004. He then moved as research group leader to ETH Zurich , Switzerland in 2007. From 2013 till 2017, he was a Research Group Leader at the Fraunhofer Society , Dresden, Germany. Since 2017, he is a professor for biomedical databases and image processing and information mining at TU Dresden Germany. In 2021 he took over a function of beeing a Group Leader of Cybersecurity Team Leader of Sax.CERT at the State of Saxony .

Research

Kozak has focused his research on large, and heterogeneous data management, with particular focus on methods from Relational Databases, Intuitive Software Frameworks, Responsive Modelling, AI (KNN; SVM and general machine learning). His success in databases are: eDOC Plattform at Max Planck Society, LIMS System, LMS 4.0, and AIMS: Kozak has design AIMS Database. Data integration is a crucial issue in the environments of heterogeneous medical or production data sources. First, there are heterogeneous data types and formats located in different databases, which implies to solve data integration challenges as a prerequisite for gaining useful information and knowledge based on appropriate analytical methods. The Advanced Information Management System (AIMS) was especially developed for this purpose. AIMS enables to request data from different locations as a routine using web user interfaces.

AIMS generates a dashboard and a report that helps the operating staff to prepare the processes through the automatic output of the associated tools and instruments. Moreover, the user can use this report to prove his preparations for the procedure and to comprehensively explain the operation to the enduser. AIMS presents an architecture which implements data integration in institutions from the production, cybersecurity and patient data. AIMS integrates databases without any changes to the individual databases (Structured Query Language, SQL) database, software backend, Application Programming Interfaces API, Frontend) nor any need to maintain another database. The solution combines database technology and a wrapper layer known from Extraction Transformation Loading systems and brings it to SQL Database, WEB API (backend) layer, Interface layer (Rest API) and frontend. It also provides semantic integration through connection mechanism between data elements. The solution allows for integration of multidomain data in one technological framework: data management platform and implementation of analytical methods in one end-user environment. Data storage in AIMS offers a highly scalable web storage service that uses cumulative digital objects rather than blocks or files. Object storage typically stores data, along with metadata that identifies and describes the content. For metadata management and automated quality control and data fusion (ETL Processes) a data consistency model (AIMS metamodel) is used to enable eventual consistency for updates or deletes to existing objects.

Kozak has applied AIMS to Domains: Cybersecurity, Medicine (MS, Neurology, Cardiology, Dental), Industry.

Awards and honors

• 2016 - visiting professor, University of Warsaw
• 2021 - visiting professor, ETH Zürich
• 2021 - Award For Excellent Innovation in Digital Stomatology
• 2019 - Innovation Award for lndustr 4.0 in Glass Production
• 2019 - Lab Management: Lab Daten für KI – Fraunhofer Zentralle München
• 2019 - ThermoFischer & IBM
• 2007 - Award for Young Life Scientists in HCS/HCA, High Content Analysis
• 2012 - SystemX award for setting up Computational Analysis Service at ETH Zurich

Bibliography

Selected articles

• Labuda N, Lepa T, Labuda M, Kozak K: Medical 4.0: Medical Data Ready for Deep and Machine Learning. J Bioanal Biomed 2017 9: 283, Vol 9(6) (2017)
• Kittler R, Pelletier L, Heninger A.K, Slabicki M., Theis M, Miroslaw L, Poser I, Lawo S, Grabner H, Kozak K, Wagner J, Surendranath V, Richter C, Bowen W, Jackson A.L, Habermann B., Hyman A.A, Buchholz F: Genome-scale RNAi profiling of cell division in human tissue culture cells, November 2007; DOI: 10.1038/ncb1659, Nature .
• Kittler R., V. Surendranath, A.K. Heninger, M. Slabicki, M. Theis, G. Putz, K. Franke, A. Cardareli, H. Grabner, K. Kozak, J.Wagner, E. Rees, B. Korn, C. Frenzel, C. Sachse, B. Soennichsen, J. Guo, J. Schelter, J. Burchard, P. S. Linsley, A.L. Jackson, B. Hebermann, F. Buchholz: Genome-wide resources of endoribonuclease-prepared short interfering RNAs for specific loss-of-function studies. Nature Methods , 2007
• Schröder T.A., Maiwald M. , Reinicke A. , Teicher U., Seidel A., Schmidt T. , Ihlenfeldt S., Kozak K., Pradel W., Lauer G. and Achour,A.B.: A Holistic Approach for the Identification of Success Factors in Secondary Cleft Osteoplasty, Journal of Personalized Medicine 12(3):506, 2022