The aim of this doctoral study is the exploitation and extension of graph-theoretical methods in tasks related to machine learning. The major focus of the study will be multi-relational networks from different resources, such as
biomedical, social, stemming from textual analysis or time-dependent graphs. In the age of big data, such networks contain latent knowledge that needs to be extracted, to provide insights for a plethora of practical and theoretical problems.
The field of multi-relational network analysis based on classical machine learning, as well as the state-of-the-art deep learning, methodologies is currently very active. The main reasons for this being the numerous research paths available and the impact that breakthroughs have on practical applications and multiple research fields in general. The purpose of the proposed research is the study of existing methodologies, leading to the creation of novel ones through the combination of ideas from different theoretical fields, aiming
to provide solutions to practical applications, while promoting research on this promising field.