Smart Data & Knowledge Services

Effective cooperation between partners in complex processes requires effective, correct and unambiguous communication. Mutual understanding between people is based on common language, shared professional knowledge and a common cultural background. Only this understanding makes it possible for people to act meaningfully and purposefully with exchanged information.

Computers lack this basis. The application of semantic technologies aims to remedy this situation: By formalizing and modeling shared concepts, a communication basis is created that enables mutual understanding for computers and thus the possibility of correct operationalization even in open application domains.

The crucial point is to enable the computer to deal correctly with the content of new information that is still unknown at the time of programming. Therefore, in all considerations of formalized knowledge representation and common vocabularies and ontologies, the dynamic evolution and growth over time must be taken into account.

Our topic area investigates representation formalisms suitable for such purposes: the Resource Description Framework RDF as a universal data representation and exchange format and other graph-oriented representation formalisms; schema languages and modeling formats such as RDFS, SKOS, OWL for modeling and logical processing of conceptual worlds and ontologies; reasoning mechanisms for testing and inferring new knowledge from existing structures - such languages and tools allow to capture and process content and relationships in a computer-adequate form.
The bridge between the formalisms and the data collected in reality is built by the various approaches of pattern recognition and machine learning: these solutions from other topic fields help to correctly interpret the data in concrete use cases and to transform the findings into formal representations.

When information is formally represented and/or semantically annotated, application-specific computer-aided assistance systems can be realized in a variety of ways: Knowledge repositories make extensive knowledge accessible to computers and humans (see also the use case “Corporate Memory” in the neighboring topic field), prediction and evaluation systems assess current and future situations and recommend actions; distributed data mining on semantically annotated data enables new insights even in complex processes.

Relevant keywords are

• Formal and graph-based knowledge representation
• Ontologies and ontology management
• Data integration and distributed data mining
• Linked Open Data and open knowledge structures
• Process-oriented knowledge management in production

Current projects use Applied Semantic Technologies in the field of industrial manufacturing, but especially in the open, complex and dynamic application domain of agriculture, where collaborating partners and modern machinery are confronted with the complex interactions of biological systems in a truly open environment with multiple external influences and yet are expected to secure food and raw material supply on a daily basis.

Selected projects:

• iGreen – Public-Private Knowledge Management
• AGATA – Analyse großer Datenmengen in Verarbeitungsprozessen
• PRO-OPT – Big Data Produktionsoptimierung in Smart Ecosystems
• GeoBox – Standardisierung der Geobox-Infrastruktur - Dezentrale Datenhaltung und regionale Vernetzung
  (Bundesministerium für Ernährung und Landwirtschaft)
• SDSD – Smarte Daten, Smarte Dienste. Landwirtschaftliche Datendrehscheibe für effiziente ressourcenschonende Prozesse
  (Bundesministerium für Ernährung und Landwirtschaft)
• SimLearn – Vervollständigung von Trainingsdaten durch iterativ lernende Simulation (BMBF)