Felipe Vargas 

Date/Time: 14 July 2023, 11:00 

Location: D2.2.094 

Abstract 

Knowledge Graphs (KGs) have gained attention as a data structure to handle disparate datasets

that contain numerical information of several individual observations in diverse domains (e.g.

agriculture, biomedical, environmental, social). Semantic Web (SW) technologies are suitable to

represent taxonomic knowledge about these KGs. However, cases that do not fall in this category

such as numerical relationships (e.g., algebraic operations or unit conversions), which can

enrich the KG data, are poorly represented. An intuitive example of a numerical relationship is

the Body Mass Index (BMI) of a person, which can be derived from their current weight and

height to enrich the initial KG. Similarly, the Vapour Pressure Deficit (VPD) in the atmosphere

can be derived from the air temperature and the air relative humidity. While experts are aware

of such mathematical formulas, most of them are performed in adhoc programming languages

limiting their reusability and reproducibility. In this thesis we explore the different Semantic

Web approaches that allow us to represent and compute these kinds of numerical relationships.

We identify some limitations of the current approaches in terms of representation, computing

methods and expressivity. To fill these gaps, we propose a Semantic-Web-based framework with

the following purposes: (i) represent the mathematical formulas conforming to LOD and FAIR

principles, in order to gain in adoption and reproducibility; (ii) on-demand execution of the

numerical relationships considering that materialisation is infeasible for large and heterogenous

KGs; (iii) express mathematical formulas using as inputs and outputs KG data in form of quantity

values to exploit semantic resources and metadata (e.g., unit ontologies); (iv) allow aggregations

within the mathematical formulas taking into account that most of this numerical data is

multi-scale. In our ongoing research we are evaluating the framework on KGs from the

agriculture and plant phenomics domain, where this thesis is carried out, as well as from more

traditional Semantic Web KGs such as DBpedia.