In recent years, we have witnessed the proliferation of knowledge graphs (KG) in various domains, aiming to support applications like question answering, recommendations, etc. A frequent task when integrating knowledge from different KGs is to find which subgraphs refer to the same real-world entity, a task largely known as the Entity Alignment. Recently, embedding methods have been used for entity alignment tasks, that learn a vector-space representation of entities which preserves their similarity in the original KGs. A wide variety of supervised, unsupervised, and semi-supervised methods have been proposed that exploit both factual (attribute based) and structural information (relation based) of entities in the KGs. Still, a quantitative assessment of their strengths and weaknesses in real-world KGs according to different performance metrics and KG characteristics is missing from the literature. In this work, we conduct the first meta-level analysis of popular embedding methods for entity alignment, based on a statistically sound methodology. Our analysis reveals statistically significant correlations of different embedding methods with various meta-features extracted by KGs and rank them in a statistically significant way according to their effectiveness across all real-world KGs of our testbed. Finally, we study interesting trade-offs in terms of methods’ effectiveness and efficiency.

Spatial Data Infrastructures (SDIs) are a key asset for Europe. This paper concentrates on unsolved issues in SDIs in Europe related to the management of semantic heterogeneities. It studies contributions and competences from two communities in this field: cartographers, authoritative data providers, and geographic information scientists on the one hand, and computer scientists working on the Web of Data on the other. During several workshops organized by the EuroSDR and Eurogeographics organizations, the authors analyzed their complementarity and discovered reasons for the difficult collaboration between these communities. They have different and sometimes conflicting perspectives on what successful SDIs should look like, as well as on priorities. We developed a proposal to integrate both perspectives, which is centered on the elaboration of an open European Geographical Knowledge Graph. Its structure reuses results from the literature on geographical information ontologies. It is associated with a multifaceted roadmap addressing interrelated aspects of SDIs.

We present a study of the evolution of the political landscape during the 2015 and 2019 presidential elections in Argentina, based on data obtained from the micro-blogging platform Twitter. We build a semantic network based on the hashtags used by all the users following at least one of the main candidates. With this network we can detect the topics that are discussed in the society. At a difference with most studies of opinion on social media, we do not choose the topics a priori, they emerge from the community structure of the semantic network instead. We assign to each user a dynamical topic vector which measures the evolution of her/his opinion in this space and allows us to monitor the similarities and differences among groups of supporters of different candidates. Our results show that the method is able to detect the dynamics of formation of opinion on different topics and, in particular, it can capture the reshaping of the political opinion landscape which has led to the inversion of result between the two rounds of 2015 election.

Data sharing in the European Union (EU) has gained new momentum, among others for machine learning (ML) and artificial intelligence (AI) training purposes. By enabling models’ training whilst preserving the privacy of data, Privacy Enhancing Technologies (PETs) have therefore gained popularity, especially among policymakers. So far, computer science research has focused on advancing state-of-the-art privacy engineering and exploring trade-offs between privacy and accuracy. Meanwhile, legal scholarship began investigating the challenges arising therefrom. Yet, few works have delved into the fairness implications of PETs. Further research is essential to both prevent the propagation of bias and discrimination and to limit the accumulation of market power within very few economic entities suitable to undermine fair competition and consumer rights. In our work, we will address this knowledge gap by adopting a legal and computer science point of view. After scoping our understanding of possible unfair sides of PETs based on technical and socio-legal understandings of fairness (Section 2), we provide an overview of PETs mostly relevant for ML and AI training (Section 3). We then discuss fairness-related challenges arising from their use (Section 4) and we suggest possible technical and regulatory (e.g., impact assessment, new rights) solutions to address the shortcomings identified (Section 5). We finally provide conclusions and ideas for future research (Section 6).