Geolocation of Cultural Heritage using Multi-View Knowledge Graph Embedding

TL;DR

This paper introduces a framework for enriching cultural heritage knowledge graphs with geographical data and a multi-view learning model to estimate distances between entities, improving geolocation accuracy.

Contribution

It presents a novel method for integrating multi-source knowledge and multi-view learning to geolocate cultural heritage entities within knowledge graphs.

Findings

Enhanced geolocation accuracy for cultural heritage entities

Effective integration of multi-source data into knowledge graphs

Improved estimation of relative distances between entities

Abstract

Knowledge Graphs (KGs) have proven to be a reliable way of structuring data. They can provide a rich source of contextual information about cultural heritage collections. However, cultural heritage KGs are far from being complete. They are often missing important attributes such as geographical location, especially for sculptures and mobile or indoor entities such as paintings. In this paper, we first present a framework for ingesting knowledge about tangible cultural heritage entities from various data sources and their connected multi-hop knowledge into a geolocalized KG. Secondly, we propose a multi-view learning model for estimating the relative distance between a given pair of cultural heritage entities, based on the geographical as well as the knowledge connections of the entities.