🧰Tools and resources exploited
Upon the suggestion of our Tutor, some tools listed below were used in the implementation of this project.
ChatGtp was used in the first phase of the project where it was requested to have a general understanding of the bias under consideration. We asked Chat GPT to provide us with a definition of the bias, some possible scenarios in which it may occur, a user story made out of one chosen scenario and possible classes and properties to build an ontology.
As an alternative to gtp chat, Gemini AI was used for the same tasks explained earlier. This was necessary where the answers provided by GPT were found to be unsatisfactory or otherwise unclear and irrelevant.
As a frame-based ontological resource collecting linguistic resources from different domains Framester was used to find semantic frames to align with the key concepts defined for each bias and possibly to use them to build classes for the ontology.
Where no suitable semantic frames could be found, resources contained in the ontology DBpedia were used as an alternative, both to create some classes and properties of ontologies.
Quokka is a tool used for QUerying Ontological resources and Knowledge bases and was used for extracting semantic frames and resources from Dbpedia and Framester. For each bias we chose the keyword concepts representing them and we put those within the Quokka Interface. The tool then provided a list of possible frames and resources related to our concepts. From that list we chose the most suitable for developing classes of each bias ontology.
Graffoo is a Graphical Framework for OWL Ontologies that we used to represent graphically with diagrams the classes and properties of each bias ontology. Thanks to this tool it is possible to display the logical relationships between elements of an ontology in a manner that is relatively straightforward to understand.
The Ontology Design Patterns website has been an essential point in the development of the project because it is from here that the conceptual reference models were drawn to build each bias ontology. In particular, the Content ODPs section was the one mainly consulted and here we chose the different design patterns that were most suitable, especially concerning the properties to be used.
This is the main tool used for visualizing graphically the classes, properties and individuals of each bias ontology. The .owl file of the ontology was used for this task.
WebOWL services is another tool for visualizing an ontology used as an alternative to OWLGrEd.
Protégé is a free, open-source ontology editor used for modeling classes and properties of each bias ontology. Every bias ontology was first developed in a separate module and then at the end each module was put in a closure module to create the final Cognitive Bias Ontologies. All ontologies have been saved in a .OWL file format. Moreover, with a new integrated plugin, it was possible to perform SPARQL on some ontologies.
Last updated