It is Time For a Language That Represents Behaviour.

June 13, 2019 by Stefan Summesberger

SEMANTiCS 2019 Keynote Speaker Andreas Harth is a professor of information systems at Friedrich-Alexander-University Erlangen-Nuremberg and Department Leader at Fraunhofer IIS-SCS. Prior to moving to Nuremberg, Andreas worked at the Karlsruhe Institute of Technology and at the Digital Enterprise Research Institute in Galway, Ireland. In this interview he talks about the behaviour of sensors and other user agents that provide the connection from the virtual to the physical world as well as the necessity for innovation in this area

Your keynote is entitled From Representing Knowledge to Representing Behaviour. This title stimulates the imagination. In which direction goes your research?

The model I work with is based on web architecture. Servers manage resources, and user agents are able to interact with resources managed by servers. Resources could be sensors or actuators that provide the connection from the virtual to the physical world.

One the one hand, the behaviour of the user agents can be reflected in the sequence of change operations the user agent carries out on resources. On the other hand, a user agent could determine the behaviour of other user agents via observing the changes of resources over time.

Check out amazing use cases and projects in the fields of Knowledge Graphs, Semantics and AI - Register for SEMANTiCS 2019!

You say, that “Smart Homes and Smart Factories, where the physical and virtual worlds are becoming connected via sensors and actuators, require languages that allow for the representation of behaviour.” What are the greatest challenges there?

For representing knowledge, mainly static knowledge, the Semantic Web community has developed languages, such as RDF, RDFS and OWL, that are very successful. However, such languages fall short in scenarios that require knowledge about change over time, i.e., workflows, processes, behaviour.

Now, the challenge is to find a language that allows for the representation of behaviour. With such behaviour descriptions, user agents could either execute behaviour or observe behaviour in the environment reflected in the resources managed by servers.

What are the most recent breakthroughs at the cross section of the representation of  knowledge and behaviour?

The last breakthroughs in the area happened about a decade ago. Many of the underlying concepts are even older. On the bright side, I think the topic is still relevant, and there is ample space for research that can generate impact, given the current development around connected devices.

The areas touched by the statement in the previous question also deeply reach into personal and critical domains. What is your approach to balancing risk and technical progress?

The systems I work on should provide the possibility to store data locally and make decisions locally, without central control. I hope that such decentralised approaches provide enough resilience and retain enough personal freedom to avoid a dystopian outcome.

You are a SEMANTiCS Veteran. What do you especially look forward to at SEMANTiCS 2019?

I look forward to hearing about the many scenarios where semantic technologies, in particular knowledge graphs, are successfully applied in practice.

Do you want to learn more about Knowledge Management, the Semantic Web and Linked Data? Register for SEMANTiCS 2019!


The annual SEMANTiCS conference is the meeting place for professionals who make semantic computing work, and understand its benefits and know its limitations. Every year, SEMANTiCS attracts information managers, IT-architects, software engineers, and researchers, from organisations ranging from NPOs, universities, public administrations to the largest companies in the world.