Semantic methods & formal ontologies
The main insight is that ontologies, models and specifications can be seens as heterogeneous modular logical theories in some logical system, formalised as institution. For complex systems, an early formal model can checked early for consistency, resulting in possible design changes – instead of doing this later (which means more involved and costly) in the development. We have successfully applied this to ontologies and specifications.
The specification and modeling language DOL ("Distributed Ontology, Model and Specification Lnaguage") has been developed within the OMG, led by Prof. Mossakowski. We have built a community around this topic and linked it to other communities. We develop a suitable language with formal semantics, as well as proof methods and tools, among them the Heterogeneous Tool Set. More...
Analogous to human information processing that integrates more rational, slow reasoning and more holistic and intuitive, faster information processing, neural-symbolic learning approaches integrate logical reasoning with neural networks. In this way, expert background knowledge (in our group, mainly ontologies and their logical axioms) can be used by neural networks in order learn both from data and axioms, thus improving performance while simultaneously reducing the amount of data that is needed for training.
While in many neural-symbolic approaches, logical and neural modules cooperate, we focus on tightly integrated approaches like conceptors and logical neural networks, because these are most promising for a new generation of artificial intelligence systems. More...
Modeling for electrical grids and renewable energies
The transition to renewable enegeries leads to challenges for the electrical grid (which will become a smart grid) and the coordination of energy production and concumption - consumers and producers will merge into prosumers. Modeling languages and tools can play a role here, in order to improve design, reliability, testing etc. of complex energy systems. We are involved in the development of the Open Energy Platform, the Open Energy Ontology and the Open Energy Knowledge Graph. We use deep learning and neuro-symbolic methods e.g. for reasoning on the knowledge graph and for complexity reduction of energy models. More... (in German)
Qualitative spatio-temporal constraint satisfaction problems
Qualitative spatio-temporal is a constraint satisfaction problem with infinite domains. Compared with quantitative methods, it offers serveral advantages: on the one hand, qualitative calculi are often more efficient, on the other hand, they come closer to human orientation in space quantitative methods. Some qualitative spatial calculi even empirically have shown to be cognitively adequate.
Traditionally, qualitative spatio-temporal calculi are modeled as relation algebras. In the last few years there has been a paradigm shift result towards the use of various methods for solving constraint problems. Problems in the area of qualitative spatial inference are often NP-hard. Therefore, often approximation algorithms are used.
Moreover, an important problem is that there has been a profileration of many calculi, varying in topic (mereotopology, absolute orientation, relative orientation) and granularity. An research question is how to heterogeneously combine these calculi, such that problems can be modeled using different calculi in an integrated and coherent way. More...