Strategic COMET project dasRES (2015-2018)

Strategic Research on Industrial Data Analysis Systems

Aims / Research Topics

The aim of this strategic project is to provide fundamental methodological building blocks necessary for robust, efficient, comprehensive and interpretable modelling and analysis of data. In particular we do research with respect to the following objectives:

  • Causal inference in the non-time series domain,
  • Learning multi-task features using deep neural architectures,
  • Probabilistic logics for knowledge representation and reasoning methods that will support an incremental industrial knowledge discovery process,
  • Computationally models for modelling uncertainty, and
  • Development of industrial feasible machine learning methods which allow analysis close to real time for industrial relevant applications.

Scientific Cooperation

  • PhD cooperation with JKU-FLLL (Johannes Kepler University Linz / Department of Knowledge-Based Mathematical Systems) on the topic of transfer learning
  • PhD cooperation with JKU-FAW (Johannes Kepler University Linz / Institute for Application Oriented Knowledge Processing) on the topic of automated data quality engineering
  • PhD cooperation with the Alfréd Rényi Institute in Hungary on the topic of semantic matching and reasoning strategies (with applications in the human resources sector)
  • Master students cooperation with KTH Sweden (School of Information and Communication Technology) on the topic of data provenance and traceability
  • Master student cooperation with JKU-FAW (Johannes Kepler University Linz / Institute for Application Oriented Knowledge Processing) on the topic of data quality assessment for NoSQL databases

Selected Publications

  • Robust estimation of correlations as basis for causal discovery [TGG15]
  • Multi-domain transfer component analysis for domain generalization [GBS+15, GBS+17]
  • Transfer learning for process analytical chemistry [LMN+15, MNP+15a, MNP+15b, MN17]
  • A novel method for domain-invariant representation learning [ZLS+17]
  • A novel approach for system identification for MPC taking into account modelling uncertainty by adaptively choosing the identification horizon [Sob17]
  • Efficient and robust median-of-means algorithms for location and regression as the basis for fast learning in big data environments [KT17]
  • Methods for online transfer learning [GCN16]
  • Combining relational and NoSQL database systems for processing sensor data [SLF15]
  • Stochastic stability analysis of perturbed learning automata [Cha17]
  • Automated data quality monitoring as the basis for meaningful knowledge representation [EW17a, EW17b]
  • A method to reason about fault detection to generate interpretable alarm lists [KT17]
  • Automated knowledge base management: a survey [Mar15]
  • A novel approach for fuzzy aggregation of semantic similarity measures [Mar16a]
  • A smart approach for matching, learning and querying information [MPS16]

Funding Partner

This project is subsidized in the frame of COMET – Competence Centers for Excellent Technologies by BMVIT, BMDW, State of Upper Austria and its scientific partners. The COMET program is handeled by FFG.


Bernhard Freudenthaler

Freudenthaler Bernhard

Area Manager Data Science
Phone: +43 50 343 850