INK: knowledge graph embeddings for node classification

Proceedings Paper Computer Science, Artificial Intelligence

FinMatcher at FinSim-2: Hypernym Detection in the Financial Services Domain using Knowledge Graphs

Jan Portisch et al.

Summary: This paper introduces the application of the FinMatcher system in the FinSim 2021 shared task, where the goal is to find the most relevant concept from a set of labels in the financial services domain. The system utilizes three publicly available knowledge graphs to generate features and predict the closest hypernym using a neural classifier.

WEB CONFERENCE 2021: COMPANION OF THE WORLD WIDE WEB CONFERENCE (WWW 2021) (2021)

Add to Collection

Proceedings Paper Computer Science, Artificial Intelligence

Walk Extraction Strategies for Node Embeddings with RDF2Vec in Knowledge Graphs

Bram Steenwinckel et al.

Summary: Knowledge Graphs require specialized techniques to be compatible with data mining techniques due to their symbolic nature. Although the Weisfeiler-Lehman kernel was originally proposed to improve the quality of node representations, our study shows limited improvement in walk embeddings within a single Knowledge Graph. Instead, we explore five alternative strategies to supplement basic random walks and demonstrate the continued relevance of research in this field for node classification tasks through comparisons on various benchmark datasets.

DATABASE AND EXPERT SYSTEMS APPLICATIONS - DEXA 2021 WORKSHOPS (2021)

Add to Collection

Article Computer Science, Artificial Intelligence