About Me and You - iTelos - A cooperative methodology for diversity-aware Knowledge Graphs

The representation of information through digital data has changed many times over the years, as has the way we interpret and use the data. Today, the increased use of AI technologies has led to an impressive amount of online data that needs to express the information it contains in a very precise and concrete way, for two main reasons. (i) To concretely express the information diversity of the geographical, social or even personal context for which the data was created, and (ii) to reduce the cost of data reuse, which is essential to support the development of data-based applications. However, the data currently available is not suitable to fully satisfy such requirements. Most of the time, data reuse is based on data transformation to better fit the purpose to be fulfilled, but the original diversity of information that the data carries is lost. This leads to a distortion of the reality represented by the reused data. A concrete case is the well-known AI services, which collect data from different local contexts, but transform this data to fit a unique, usually Western and business-oriented data model. This data reuse problem is exacerbated by a lack of data representation models that can concretely represent local information diversity. And a lack of methodologies to support data reuse. We propose iTelos, namely a methodology for data reuse based on the production, integration and distribution of diversity-aware Knowledge Graphs (KGs). iTelos is based on three main processes. The first aims at representing a new type of data, based on the Entity Base data model, capable of concretely representing the information diversity of a context. The second is a KG completion process, for the integration of diversity-aware data. The third one is a data distribution process, which aims to enhance the reuse of diversity-aware data by supporting the collaboration between those who produce data and those who are interested in using it, within a data distribution network called LiveData. The iTelos methodology supports both expert users in the creation of diversity-aware data and non-expert users in the development of data-based applications by providing them with the necessary data and support, and ultimately by establishing collaboration with expert users. The methodology has been tested in the Knowledge Graph Engineering (KGE) Masters course at the University of Trento (Italy), where students used iTelos to develop KGs suitable for solving several real-world problems. The first version of the LiveData network has also been developed by the University of Trento (Italy) and the National University of Mongolia (Ulaanbaatar, Mongolia).