Python Based Imperative Representation of Knowledge (Pyirk)
Pyirk is an experimental framework for imperative knowledge representation. I is basically a collection of classes and functions which facilitates the construction of and the interaction with knowledge graphs.
The framewok is structurally inspired by
-
allow for statements about statements by means of qualifiers
provide an RDF export and SPARQL interface
-
model the content of scientific contributions, instead of mostly the metadata
Suggested Upper Merged Ontology (SUMO)
modelling of higher order statements
However, pyirk is a pure python framework and aims to be intuitively usable without prior familiarity with knowledge engineering techniques such as OWL, but instead requiring only some understanding of programming.
Motivation
Humankind is faced with an growing amount of knowledge available (in principle) and this growth is accelerating by itself. However, this knowledge often cannot be applied to solve actual problems because it is stored somewhere out of reach and often also distributed over several documents which often are not entirely consistent.
While library catalogues, full text search engines and similar facilities help a lot in the procurement of relevant knowledge, they underlie some fundamental limitations. In 2001 the term Semantic Web was introduced to refer to (existing and future) technologies which allow a machine – in some sense – to understand information instead of just storing and processing it. However, even after more than two decades, apart from some niche fields such semantic technologies yet have not had a mayor impact on science and societies, especially if compared with technologies such as machine learning.
One reason is that the available technologies like OWL (based on so called description logic) are widely considered to be hard to grasp by non-specialists. The current mainstream approach of formal knowledge representation thus consists the cooperation between domain-experts and knowledge engineers to create useful knowledge bases for specific domains.
In contrast, Pyirk aims to enable domain-experts themselves to create such knowledge bases, by providing an interface in the widespread Python programming language, which poses a significantly lower barrier compared to the special purpose OWL.
Pyirk also aims to allow for much greater expressive power (e.g. allowing higher order logic statements) than most other approaches, despite the computational consequences.
Status
The whole Pyirk project and even much more this documentation is currently still under development and should be considered as incomplete and only partially functional. Nevertheless some usefulness cannot be excluded.
User Documentation
Information regarding the local usage of pyirk.
Developer Documentation
Information regarding actively contributing to pyirk.
CLI
More
Note
The following links are not yet functional.