The PRAXICON is a conceptual knowledge base in which concepts have both symbolic (e.g. natural language) and corresponding sensorimotor representations, while associations among them are pragmatic in nature and recursive. The resource has been developed to allow artificial agents/systems:
- to tie concepts/words of different levels of abstraction to their sensorimotor instantiations (catering thus for disambiguation), and
- to untie sensorimotor representations from their physical specificities correlating them to conceptual structures of different levels of abstraction (catering thus for intentionality indication).
The PRAXICON database is a 3rd normal form database that stores information on embodied concepts. The design of the database ensures that any application that uses the PRAXICON knowledge base and its tools will have easy access to the stored data, with minimum overhead. To this end, we have included some features that are not currently used by the POETICON project applications, but will be useful for future extensions of the PRAXICON and use in different applications (e.g. in machine translation).
The core entity of the database is the Concept table. It connects concepts/words of different levels of abstraction to their sensorimotor and visual instantiations. Each concept has a linguistic representation and can optionally have a motoric and/or a visual representation.
Each Concept can be member of one or more Relations. Relations usually comprise of a triplet having the following form: "Concept A – Relation – Concept B". This means that Concept A (left relation argument) is related with some Relation to Concept B. These Relations are always bidirectional, so the aforementioned triplet also stands for Concept B is related (with the reverse relation) with Concept A. These triplets are the simplest form of a Relation.
This simple form of a Relation cannot by itself express the perplexity of the world we live in. We need to be able to combine simple Relations into more complex structures. These are called Relation Sets. A Relation Set can contain one or more Relations bundled together as one entity.
In addition, a Relation Set can be at the left or right side of a Relation, thus creating a recursive schema. More details on how Relations interact with Relation Sets and Concepts can be found in the documentation.
There are two possible ways to get the PraxiconDB:
-
Get the latest stable release from https://github.com/CSRI/PraxiconDB/tags.
-
Get the latest development version by creating a local fork of the project on your Git account and clone it locally at https://github.com/CSRI/PraxiconDB.git.
For installation and usage instructions please refer to the documentation
in the manuals under docs/.
If you spot any bugs, please report them to the GitHub's issue tracker: https://github.com/CSRI/PraxiconDB/issues.
If you would like to contribute to the project, just fork it, do your magic and send pull requests with your changes.