Dr.-Ing. Tanja Auge

I am a PostDoc at the chair for Data Engineering at the University of Regensburg. Since October 2022 I am focusing on reconstructability/reproducibility and plausibility in the field of data engineering. For this, I am cooperating with researches from different fields at the University of Regensburg as well as international colleges from the US, Germany or Austria.

Before joining University of Regensburg, I was a PhD student at the University of Rostock advised by Andraes Heuer at the Data Research Group. My PhD thesis is available here. Prior to that, I studied Computer Science (M.Sc.) and Mathematics (M.Sc. and B.Sc.) at the University of Hamburg and the University of Rostock.

In my PhD thesis I combined the Chase -- a family of algorithms for database transformation -- with data provenance and additional annotations to compute an anonymized (minimal) sub-database of an original (research) database for a given evaluation query. To ensure the reproducibility, replicability or plausibility of a query result, the evaluations performed on the original database must also be feasible on the reconstructed sub-database. For this purpose I used a version of Chase&Backchase, an extension of Chase.

Projects

• Research Data Management (Website)
• ProSA: Provenance Management using Schema mappings with Annotations (Website) 

Research interests

My interests include a variety of subjects, such as

• provenance, in particular data provenance,
• research data management,
• Chase algorithm,
• schema evolution and data updates, and.
• data provenance for AI.

For me, access to data and its analysis in science and society is the cornerstone of free and sound research. The open, FAIR and privacy-compliant provision of this data is therefore one of our most important basic obligations. My goal in provenance research is to ensure the traceability of a (published) result throughout the entire data science life cycle back to its (possibly physical) source. Especially for machine-generated analysis results, it is particularly important to know where the data used comes from, as it is often (pre-)processed in several stages and may originate from unclear sources. Data from the internet or social media in particular often has a very short half-life and is then no longer traceable. Possible boundary conditions that must be taken into account here include the evolution of the databases at schema and data level, the database type and compliance with privacy aspects. Areas of application for this include research data management and various data engineering, data science or AI applications. With my research, I would like to contribute to making scientific, but also economic processes and evaluations more transparent and thus reduce or prevent errors in data processing as well as ambiguities in their interpretation, hallucinations (of LLMs) and the like. 

Publications (extraction)

• L. Waltersdorfer, D. Hausler, T. Auge:
  Provenance Question-based Selection for AI Transparency and Accountable AI Governance.
  (Accepted for AIGOV@AAAI 2025)
• T. Auge, S. Genehr, M. Klettke, F. Krüger, M. Schröder:
  Towards dimensions and granularity in a unified workflow and data provenance framework.
  (Presented at LWDA 2024)
• T. Auge, L. Waltersdorfer, E. Michels, S. Feistel, S. Jürgensmann, F. J. Ekaputra, M. Klettke:
  Towards an Integrated Provenance Framework - A Scenario for Marine Data. 
  TaPP@EuroS&P Workshops, 2024 (DOI)
• T. Auge, F. J. Ekaputra, S. Feistel, S. Jürgensmann, M. Klettke, L. Waltersdorfer:
  Challenges of Tracking Provenance in Marine Data. 
  IMDIS, 2024 (pdf)
• S. Diemt, T. Auge:
  Anforderungsanalyse für ein universitätsweites nutzerorientiertes Forschungsdatenmanagementsystem basierend auf einer Nutzerumfrage. Softwaretechnik-Trends 44(1), 2024 (pdf)
• M. L. Möller, D. Hausler, S. Strasser, T. Auge and M. Klettke: 
  Heterogeneity in NoSQL Databases: Challenges of Handling schema-less Data. 
  LWDA, 2023 (pdf)

• T. Auge, G. Bali, M. Klettke, B. Ludäscher, W. Söldner, S. Weishäupl, T. Wettig:
  Provenance for Lattice QCD workflows.
  TaPP@WWW, 2023 (DOI)

• T. Auge:
  ProSA - A provenance system for reproducing query results.
  TaPP@WWW, 2023 (DOI)

• T. Auge:
  Provenance Management unter Verwendung von Schemaabbildungen mit Annotationen.
  PhD Thesis, University of Rostock, 2023 (pdf)

• T. Auge, A. Heuer: 
  Tracing the History of the Baltic Sea Oxygen Level Evolution and Provenance for Research Data Management. 
  BTW, 2021 (DOI)

• T. Auge: 
  Extended Provenance Management for Data Science Applications. 
  PhD@VLDB, 2020 (pdf)

For further publications see dblp or the list below.

Talks

• ProSA Pipeline — Provenance conquers the CHASE. University of Illinois at Urbana-Champaign, School of Information Sciences, 2022

• Schema Evolution in Research Data. Spring Symposium Databaeses, 2024