The rapid growth of online data and the widespread use of databases have created an immense need for Knowledge Discovery methodologies. The challenge of extracting knowledge from data draws upon research in statistics, semantic web and linked data technologies, databases, language processing, data mining, data visualization, optimization, and high-performance computing, to deliver advanced data analysis and discovery solutions. Knowledge discovery is the process concerned with automatically analysing large volumes of data to derive patterns that can be considered knowledge about the data. Knowledge obtained through this process should be actionable and is in itself exploitable for further usage and discovery.  The Knowledge Discovery Unit KDU focuses on the investigation of the principles and applications behind the next generation of intelligent, meaning-oriented systems, investigating new approaches for the semantic representation, processing and analysis of large-scale heterogeneous data in context. The Unit focus lies at the intersection between unstructured and structured data sources in order to derive actionable knowledge.  The unit is cross-disciplinary and conducts high impact research across the fields of Natural Language Processing, Semantic Web, Linked Data and Data Visualization and Graph Analytics .


  • scalable state of the art infrastructures and techniques for knowledge discovery and capture across heterogeneous data sources in the domain context i.e. streaming sensor data, text and network analysis.
  • methodologies and processes for designing and evaluating interfaces for the collaborative creation, management and access of formal knowledge.
  • data visualisation techniques, processes, tools and algorithms and for scalable knowledge  discovery from heterogeneous (integrated unstructured and structured sources)
  • scalable and robust Integration of language, social, semantic and graph analysis technologies for knowledge discovery in context i.e. specific applications are finance and scientiometrics currently.