The ability to *represent knowledge about the world* and to *draw logical inferences* is one of the central components of intelligent behavior, as a consequence, reasoning components of some form are at the heart of many artificial intelligence systems.

## Research

#### Research

The KWARC research group conducts research in knowledge representation and reasoning techniques with a view towards applications in knowledge management.

We extend techniques from [formal methods](http://kwarc.info/semantics.html#fm) so that they can be used in settings where formalization is either infeasible or too costly.

We concentrate on developing techniques for marking up the [structural semantics](http://kwarc.info/semantics.html#ssem) in technical documents.

This level of markup allows for offering interesting [knowledge management services](http://kwarc.info/projects/) without forcing the author to formalize the document contents.

## Research Domain/Strategy: Corpus-based Meta-Mathematics

#### Research Domain/Strategy: Corpus-based Meta-Mathematics

In this title we have three components that warrant explanation:

S1. **Mathematics**: we use the mathematical knowledge and mathematical documents as a test

1.**Mathematics**: we use the mathematical knowledge and mathematical documents as a test

tube for our research as the knowledge and document structures are quite explicit and

overt and the content of mathematics is well-understood. Anything that has the same

properties we consider to be "mathematics" as well.

S2. **Meta**: we develop Meta-Artefacts, i.e. we

2.**Meta**: we develop Meta-Artefacts, i.e. we

* design **representation languages** (logics) that allow to talk *about* mathematical objects,

their properties, and relations,

* invent **algorithms** that analyze and transform these representations, and

* implement them in **end-user systems** that utilize both.

S3. **Corpus-based**: we do this as a natural science by looking at data (i.e. corpora of

3.**Corpus-based**: we do this as a natural science by looking at data (i.e. corpora of