Despite the availability of numerous open datasets on cultural heritage, limited research has focussed on structuring and normalising this type of data, particularly through the extraction of entities from unstructured texts. This step is crucial for enriching, analysing, and understanding these complex datasets. This study presents a procedure designed to streamline the creation of domain-specific datasets for training natural language processing models and evaluates their performance across three distinct datasets generated using this procedure. A zero-shot learning model, the Generalist and Lightweight Model for Named Entity Recognition, was assessed alongside pre-trained spaCy models on three datasets created in the framework of the European Union-funded Research Intelligence Technology for Heritage and Market Security project: one containing provenance information on artefacts from North American museums, another detailing stolen cultural goods in Romania, and a third with structured yet unclassified data on WWII-looted Polish art. Further training of spaCy models on these newly defined datasets revealed that fine-tuned models significantly outperform their non-fine-tuned counterparts, with the best results from the Transformer model fine-tuned on provenance data. This success can be largely attributed to the standardised conventions in provenance research. In contrast, the model fine-tuned on descriptive information performed poorly, likely due to extensive descriptions containing non-essential data that increased model uncertainty. This work highlights the potential of automating entity extraction to build knowledge graphs for cultural object databases, enabling advanced analytical approaches such as Network Analysis.

The current landscape for provenance researchers, cultural heritage crime analysts, and law enforcement working in the culture sector is characterised by the siloing of data across dozens of databases, resulting in fragmented and incomplete resources that must be manually correlated and validated to provide insights into cultural heritage crime. The European Union-funded Research, Intelligence, and Technology for Heritage and Market Security (RITHMS) project is developing a platform to assist law enforcement agencies across Europe in tackling the illicit trafficking of cultural goods by aggregating open, specialised, and police data from a range of sources. This article outlines one of the initial phases of data collection, which has developed 30 tailored web scrapers for the collection of data from existing databases of stolen, missing, protected, and unprovenanced cultural goods. This has resulted in the largest known non-police dataset of these and associated data objects addressing the real-world challenge of data siloes in heritage crime and provenance research. This article details the multi-step process of data collection and pre-processing that has produced the novel consolidated dataset. The mechanism for knowledge discovery developed during this project has immediate applications and has resulted in actionable intelligence for the investigation of cultural goods crimes, highlighting the value of consolidated data as a resource. This research also offers a cursory analysis of the resulting dataset, demonstrating how mining of this resource can enable new scientific insights and offer promising opportunities for intelligence-led policing of cultural heritage crimes.

Until March 2024, Ukraine’s National Agency on Corruption Prevention (NACP) was tasked with maintaining databases of individuals sanctioned for their support of Russian military action in Ukraine and monitoring works of art owned by these individuals to prevent their being sold in violation of sanctions. The present study gathered data from the NACP’s War & Sanctions and War & Art databases through a combination of web scraping and API requests. These data were then examined using measures of network analysis (including degree centrality, eigenvector centrality, and clustering coefficient), which allowed the visualisation and statistical analysis of the network of Russian elite. This research is the first that has conducted scientific analysis on these datasets. Investigation determined that the social network of individuals in the upper strata of Russian society is vast and moderately centralised, with over a dozen clusters around influential Russian oligarchs, politicians, and celebrities. It also determined that art collectors have a more central role in this network than non-collectors of art. The network exhibits a clear presence of hubs and nodes with high degree centrality and a high degree-betweenness centrality correlation, indicating that information can flow quickly and efficiently. However, the network’s dependence on a few key individuals could constitute a vulnerability, with the extraction or isolation of these figures potentially disrupting the network’s function and coordination. This research also identified inconsistencies in the imposition of sanctions, creating grey areas that could allow the continued trade of cultural goods, potentially bypassing international restrictions.

This paper delves into the issue of ‘conflict antiquities’, that is, the phenomenon of profiting from cultural goods looted in archaeological sites affected by wars or civil conflicts. The present research shows the increased number of Palmyrene funerary portraits of dubious provenance circulating on the antiquities market after the beginning of the so-called ‘Arab Spring’ in 2011. By analysing international war damage reports, the author established a chronology of the key events that have involved the ancient town of Palmyra, likely leading to the flood of portraits recently offered on sale.