Abstract
This paper presents a semantic hypermedia framework for documenting and interpreting cultural heritage artifacts, with particular attention to decontextualized and refunctionalized architectural components, through curated integration of heterogeneous sources. Built on Linked Open Data principles, it addresses two key challenges: mediating semantic complexity for domain experts, and ensuring data quality when integrating external resources. Knowledge patterns map CIDOC-CRM to a project-specific model aligned with scholarly practice, while external datasets—such as the Getty vocabularies, OpenStreetMap, Zotero, and Iconclass—are incorporated through a graded model of semantic commitment, distinguishing authority alignment, partial reuse, and full ontology adoption. External information is materialized in a knowledge graph to ensure reproducibility and long-term accessibility. By framing external linking and semantic mediation as a hypermedia design problem, the approach shows how controlled integration can support navigation, enrichment, and scholarly reuse of cultural heritage data while preserving curatorial control. We contribute a practical model for dataset curation, semantic mediation, and graph-based knowledge representation in cultural heritage hypermedia systems, developed through the design and implementation of a case study.
1 Introduction
The historical dynamics of cultural exchange and mobility in medieval Southern Italy resulted in frequent transformations of architectural elements—manifested in changes of location, function, and symbolic meaning [12]. These transformations produce complex and distributed historical evidence that is difficult to represent through conventional cataloging practices. Consequently, scholars must systematically collect and compare data on dispersed and decontextualized objects across multiple spatial and temporal scales [12]. This requirement motivates the adoption of a semantic structure that can explicitly represent evolving relationships [2].
A central problem lies in reconciling the expressive power of semantic models with the research needs of domain experts. While ontologies such as CIDOC-CRM enable rich representation of cultural heritage data [6], their complexity often exceeds not the expertise of humanists, but at least their assumptions about what limitations exist when encoding their research questions. This problem is generally well-understood by the designers of platforms for publishing knowledge in the Humanities, and among their responses is the introduction of knowledge patterns (KPs): families of parametric queries that mediate between a project-specific local model and the global ontology of the wider underlying knowledge graph [15]. While an immediate adoption case for KPs is to construct field definitions for users to input and navigate data without direct exposure to underlying semantic complexity, these can also set the rules that realize the classical paradigm of view-based data integration [10].
A second challenge concerns the integration of external resources while preserving data quality and interpretive control. Domain experts in the Humanities may report a perceived lack of curatorial control whenever a strong semantic link causes external statements to be imported and displayed as though true within the controlled research environment, or generates a hyperlink branching outside that environment. Nevertheless, the curatorial effort should depend solely on the specifics of each project and the unique knowledge it aims to provide, rather than being spent on repeated tasks only to achieve apparent completeness. For example, a project aimed at representing the history of monumental heritage should not be concerned with building a gazetteer of contemporary geography, for the sole sake of plotting objects on a map; if, however, the scholars can provide highly specialized knowledge on the topology of monuments, this should be integrated with whatever knowledge an external gazetteer can offer.
The approach is implemented in an ongoing research project dedicated to documenting liturgical furnishings in medieval Southern Italy [11]. Rather than exposing external datasets as direct links, the system incorporates them through curated workflows over the local knowledge graph. This approach is formalized as a graded model of semantic commitment, which differentiates between authority alignment, partial semantic reuse, and full ontology adoption, enabling controlled interoperability across heterogeneous sources.
Within this framework, the dataset is structured as a multi-layered knowledge graph that encodes part–whole and spatial relationships across three levels—architectural sites, ritual objects, and their components—supporting both precise historical representation and navigable hypermedia exploration.
2 Related work
Recent work has framed knowledge graphs as environments enabling relational and non-linear exploration. Mauro et al. demonstrate how semantically interlinked datasets allow cultural heritage narratives to emerge through navigation across connected entities, positioning linked data collections as hypermedia systems instead of static repositories [13].
User interaction with semantic infrastructures has likewise received increasing attention. Mulholland et al. emphasize mediation strategies that enable non-technical users to curate and explore complex datasets without direct exposure to underlying ontologies [14]. In this context, Oldman and Tanase describe the ResearchSpace (RS) platform as an environment supporting collaborative scholarly modeling through reusable KPs, which mediate between complex semantic models and domain-specific research practices [15].
Applications in digital art history further illustrate these principles. The Staccioli Digital Archive shows how knowledge graphs can feed research catalogs and exhibition contexts by connecting artworks, archival documentation, and interpretative data within a unified semantic structure [16].
Beyond the CH domain, the LED project explored how linked data ecosystems can integrate external sources using large dataset indices, while maintaining data quality through curated workflows [1]. This work highlights the necessity of balancing openness with controlled semantic integration, particularly in environments where heterogeneous datasets may imply different levels of semantic commitment.
Research on identity links in Linked Data has shown that constructs such as owl:sameAs are frequently applied beyond strict logical equivalence, leading to risks of semantic overcommitment and unreliable integration across heterogeneous datasets [5, 8, 9]. This observation motivates the need for differentiated integration strategies, as not all external links warrant the same level of semantic commitment, a principle that is operationalized in the graded model proposed in this work.
More broadly, research in Linked Open Data and knowledge graph design has emphasized the importance of governance mechanisms that preserve provenance, semantic consistency, and interpretative reliability [6, 7]. Within semantic hypermedia environments, such mechanisms function as safeguards that mitigate the risks of semantic overcommitment and uncontrolled data integration.
Building on these perspectives, this paper addresses two interrelated challenges: mediating the complexity of ontology-driven representations for domain experts and ensuring data quality in the integration of external resources. To this end, a semantic hypermedia framework is proposed, which combines knowledge-pattern–based mediation with multiple degrees of semantic commitment, enabling controlled integration and navigable interaction within cultural heritage knowledge graphs.
3 Approach
The case study driving this research aims at cataloging liturgical furnishings. The dataset is structured as a multi-layered system that integrates part–whole relationships, with the spatial context being integrated through a variety of relationships, such as the provenance of the furnishings and the history of their preservation. Liturgical furnishings can be ritual objects—such as altars or baptismal fonts—situated within architectural sites, which may themselves consist of interconnected elements like churches, abbeys or cloisters within larger monumental complexes. At a finer level, these objects are further decomposed into components, such as slabs, which may have since lost their original context and function and become artifacts in their own right, which are treated atomically by the art historian.
Each entity in the database—even a measure or a point in time—is modeled as a node within a graph, linked via CIDOC-CRM predicates and enriched with spatial, bibliographic and iconographic information, supporting both automated queries and interactive navigation. Nodes of any type can, potentially, be presented as a Web page but, to enable effective interaction with the dataset, a design decision for the catalog is to reduce the likelihood of hyperlinks leading to pages of types that are not considered to be the focus of the project. One typical example is events, such as the production of an object or the removal of part of it, both of which are represented through an event system, as mandated by CIDOC. Contributors, however, do not directly model events as such. For instance, they will want to record that an altar was produced in 1230, without explicitly representing the production event, and may wish to flag a component as lost without modeling the event that led to its removal from its original location.
The complexity of the underlying semantic ontology is therefore mediated through a project-specific local model that does not expose any part of CIDOC and is instead oriented toward navigation and use. Within RS, this mediation is implemented through reusable KPs, configurable entity templates, and collaborative workflows that support incremental contributions by multiple researchers. KPs translate semantic relations into structured interaction points, allowing users to author and traverse connections without direct engagement with the ontological schema. In this way, research questions are expressed through guided input forms that generate consistent semantic assertions while maintaining an accessible environment for non-technical collaborators.
This mediated structure shapes how the dataset is explored. Users do not encounter the ontology directly; instead, they navigate a relational knowledge space in which objects, places, materials, functions, and historical transformations appear as interconnected nodes. The semantic model operates primarily as an infrastructural layer, enabling faceted search and exploratory navigation while remaining largely invisible at the level of user interaction. KPs thus function simultaneously as mechanisms of data entry and as pathways for hypermedia traversal across the knowledge graph. Through this approach, the platform supports the full lifecycle of research data [15], from collaborative input and validation to interpretation and discovery. Interoperability is pursued through curated connectivity: external links are selectively reviewed and maintained by project leads. This design reflects a deliberate balance between openness and scholarly control, ensuring data reliability while sustaining a coherent semantic hypermedia environment.
3.1 Managing External Sources through Graded Semantic Commitment
The graded semantic model addresses the tension between interoperability with external linked data and the need to preserve scholarly control over interpretation, provenance, and data reliability within a research-oriented knowledge graph. This model describes an approach to data integration in which information is incorporated with different levels of semantic explicitness and ontological alignment. Instead of enforcing uniform formalization, it recognizes that heterogeneous datasets embody diverse modeling assumptions and degrees of curation. Semantic relations may therefore range from simple identification links to fully integrated ontological representations supporting inference. Such models enable interoperability while accommodating the practical and epistemic diversity characteristic of large knowledge-graph environments.
This strategy reflects a broader understanding of semantic integration as a graded process rather than a uniform operation. In semantic information systems, external resources differ in structure, stability, and conceptual scope, and therefore cannot be integrated under a single semantic assumption. A graded semantic model acknowledges these differences by allowing entities and relationships to be incorporated with varying levels of formal alignment and interpretative commitment. Semantic connections may range from lightweight authority references to deeper reuse of external conceptual structures, depending on the reliability of the source and the role it plays within the research context. Such an approach enables interoperability while preserving scholarly control over meaning construction and data quality within the knowledge graph.
In the context of the project at hand, this graded perspective is operationalized through a structured model that governs how external datasets are incorporated into the platform. Instead of relying on a uniform integration strategy, external resources are evaluated according to their epistemic role, stability, and compatibility with the project's conceptual framework. This results in distinct levels of semantic integration, ranging from lightweight authority alignment to partial semantic reuse and full ontological adoption. Each level defines how external identifiers are interpreted, how much external structure is preserved, and to what extent semantic assertions become part of the local knowledge graph. The following sections describe these levels of commitment, as illustrated in Figure 1, and their implementation within the RS environment.
Diagram showing three concentric levels of semantic commitment—authority alignment, partial semantic reuse, and full ontology adoption—each associated with example external resources such as Getty vocabularies, Wikidata, OpenStreetMap, Zotero, and Iconclass.
3.2 Authority-Level Alignment: Getty Vocabularies and Wikidata
The differentiated integration strategy adopted by the platform resonates with longstanding debates within Semantic Web research concerning the semantics of identity links, particularly the use of owl:sameAs and other strict equivalence assertions. As demonstrated by Halpin [8], the formal interpretation of owl:sameAs as strict logical identity often conflicts with its practical use on the Web, where links frequently express weaker relations such as similarity, correspondence, or contextual equivalence rather than true ontological sameness. Empirical analyses by Ding et al. [5] further reveal that identity links in Linked Open Data ecosystems are applied inconsistently, introducing risks of semantic overcommitment and unintended inference propagation across datasets. Similarly, Jaffri, Glaser, and Millard [9] emphasize that identity management in distributed semantic environments requires explicit strategies to balance interoperability with trust and data provenance.
Within this landscape, controlled vocabularies such as the Getty AAT and Wikidata serve as key reference points, both being predicated upon in-house metamodels—one thesaural and one ontological, respectively. This level provides a curated conceptual structure with stable identifiers and, in LOD form, semantic alignments, such as owl:sameAs, to external authorities, thereby establishing strong conceptual equivalence at the level of controlled terminology. In the RS platform, these resources are primarily used to align local entities—such as materials, types of artifacts, or techniques—with externally maintained concepts, often through identity assertions. For instance, a marble component recorded in the knowledge graph may be associated with “Rosso antico”, or taenarium, and linked via owl:sameAs to its corresponding Wikidata entity (Q3849735). This establishes a stable authority alignment, ensuring consistent terminology across datasets while maintaining interpretive control within the local research environment.
3.3 Partial Semantic Commitment: OpenStreetMap and Zotero
A second level involves partial semantic commitment, applied to datasets that provide essential contextual information but require mediation before being integrated. For spatial data, OpenStreetMap (OSM) serves as the primary georeferencing infrastructure. Scholars contribute directly to OSM by refining monument and complex boundaries according to their domain expertise. Through a subsequent georeconciliation process, entities in the project dataset are linked to their OSM counterparts via queries to Nominatim, OSM's geocoding API.1 Automated scripts subsequently retrieve coordinates and administrative hierarchies using the Nominatim and Overpass APIs and convert them into CIDOC CRM–compliant representations stored locally. Because OSM identifiers may change over time, especially if a building's polygon is redrawn from scratch, links are stabilized through linking to the corresponding Wikidata entities, cementing semantic validity across updates.
Though profoundly different in meaning, bibliographic information follows a comparable workflow using a separately curated Zotero collection [4]. Upon populating the dataset, experts select references through KPs that query a SPARQL wrapper to Zotero's API, after which post-processing routines retrieve structured publication metadata and incorporate them into the knowledge graph. In both cases, external schemas inform the data structure, yet the project maintains interpretative control through transformation and validation processes.
3.4 Full Semantic Commitment: Iconclass
The highest level of integration corresponds to full semantic commitment and is exemplified in the project by Iconclass, the standard thesaurus for iconography in art history [3]. Unlike OSM or Zotero, Iconclass functions as an authoritative conceptual system whose taxonomy, formalized using the SKOS metamodel, is incorporated as an established external structure and is not subject to local modification or collaborative editing. Iconographic concepts are integrated as stable semantic entities within the graph and linked through CRM relations that formally denote representational meaning. The project therefore relies on Iconclass as a fixed interpretative framework, ensuring consistency in iconographic classification across datasets and contributors.
The graded model transforms external linking into a hypermedia design strategy. Different resources participate in it according to their epistemic role: authority alignment, contextual enrichment, or conceptual structuring. By distinguishing levels of semantic commitment, the system maintains navigable semantic relations while safeguarding scholarly data quality. The resulting infrastructure supports collaborative enrichment without exposing researchers to the complexity or instability of external ontological systems.
3.5 Navigation with the faceted search
The differentiated integration of external resources ultimately serves a practical objective: enabling intuitive navigation of the knowledge graph through KPs. The underlying semantic structure remains largely in the background, while users explore the dataset through a faceted search interface that turns semantic relations into navigable research pathways aligned with the local model. This interface allows scholars to combine multiple parameters—such as resource type, geographic location, material, or iconographic motif—in order to identify meaningful connections across dispersed objects and contexts. To support such exploratory queries, data are structured to accommodate the simultaneous application of multiple filters for search and browsing. For instance, users may retrieve ritual objects executed in the technique opus sectile and associated with specific iconographic themes such as prophets or saints. In this framework, KPs function as mediation layers that translate complex semantic assertions into searchable dimensions.
Iconographic data play a particularly important role within this navigation model. The integration of Iconclass enables hierarchical browsing through a controlled taxonomy, while the faceted interface exposes the classification tree only up to an intermediate level, e.g. “saints or prophets” as opposed to “ox (possibly with book); symbol of St. Luke”, as illustrated in Figure 2. This design choice avoids both excessive abstraction and unnecessary granularity and it is made possible through semantic commitment with Iconclass: users are guided toward conceptually coherent iconographic categories without requiring detailed familiarity with the full Iconclass system. As a result, the semantic structure becomes not only a mechanism for data modeling but also a hypermedia interface that supports exploratory research and discovery.
4 Implementation
The project examined in this paper is implemented as a suite of templates, KPs and service wrappers that customize the ResearchSpace platform. The system combines RDF-based repositories and non-RDF web services through differentiated access and integration strategies aligned with the graded model of semantic commitment.
External resources are accessed through two main paradigms: SPARQL endpoints and non-RDF web APIs. Both are queried via SPARQL, including the use of federated queries per the W3C specification,2; however, non-RDF services returning JSON or XML content are integrated through custom SAIL wrappers3 that expose them as if they were RDF graphs. To support efficient interaction, the system distinguishes between search and lookup operations: search services identify candidate entities, while lookup services retrieve authoritative data about entities with known identifiers as a nightly maintenance operation.
Integration is implemented through the ResearchSpace Ephedra framework, which enables federation across heterogeneous sources, including opaque federation to non-RDF APIs. Custom SAIL services were developed for resources such as Nominatim and Zotero, enabling their reuse within the semantic environment. Due to platform limitations, real-time access to external services during data entry is restricted; therefore, a post-processing pipeline materializes a minimal subset of external data—such as labels, geospatial information, and bibliographic metadata—within the local knowledge graph. This achieves a functional trade-off among data persistence, reproducibility, and controlled integration while complying with external service constraints.
5 Conclusion
This paper has worked through the design and construction of a semantic hypermedia framework intended to address the complex research needs of cultural heritage scholars. The presented case study described how hierarchical modeling combined with graded semantic commitment can enable the curated integration of external resources—from authority vocabularies to fully aligned datasets—while preserving interpretive control and data quality.
The platform transforms external links into navigable structural extensions of the knowledge graph, allowing scholars to explore provenance, refunctionalization, and historical relationships without direct exposure to the underlying ontological complexity. By mediating semantic depth and selectively concealing technical structures through knowledge patterns, the approach is designed to support intuitive interaction while maintaining semantic rigor. It offers a transferable blueprint for building interoperable semantic hypermedia systems that reconcile complex data infrastructures with effective scholarly interaction. The framework has so far been developed and applied within a single ongoing case study; a systematic evaluation of its usability, scalability, and applicability to other cultural heritage projects remains a direction for future work.
While extending external linking to further open data sources is not an immediate priority, the need for doing so will be monitored to assess whether it would require introducing additional levels into the graded model. As the project's dataset is scheduled to reach critical mass by Fall 2026 and subsequently be published as Linked Open Data, a user study with art historians and other domain experts is being planned. By covering use cases in data curation and in external engagement with the platform, the study will gauge how the hypermedia rendering of data integration affects perceived information richness and trustworthiness.
Acknowledgments
The research for the present publication has been supported by the Bibliotheca Hertziana – Max Planck Institute for Art History (project No. BH-P-21-26).
Notes
1OpenStreetMap Nominatim API, https://nominatim.openstreetmap.org/.
2SPARQL 1.1 Federated Query – W3C Recommendation 21 March 2013, https://www.w3.org/TR/sparql11-federated-query/.
3Repository and SAIL Configuration, https://rdf4j.org/documentation/reference/configuration/.
Source
Imported from ACM’s structured HTML source. ACM Reference Format: Polina Voronova and Alessandro Adamou. 2026. A Graded Model of Semantic Commitment for Curated Cultural Heritage Data Integration. In 37th ACM Conference on Hypertext (HT '26), September 14--18, 2026, London, United Kingdom. ACM, New York, NY, USA 6 Pages. https://doi.org/10.1145/3800935.3830857
References
[1] Alessandro Adamou, Mathieu d'Aquin, Helen Barlow, and Simon Brown. 2014. LED: curated and crowdsourced Linked Data on Music Listening Experiences. In Proceedings of the ISWC 2014 Posters & Demonstrations Track a track within the 13th International Semantic Web Conference, ISWC 2014, Riva del Garda, Italy, October 21, 2014(CEUR Workshop Proceedings, Vol. 1272), Matthew Horridge, Marco Rospocher, and Jacco van Ossenbruggen (Eds.). CEUR-WS.org, London, United Kingdom, 93–96. https://ceur-ws.org/Vol-1272/paper74.pdf
[2] Alessandro Adamou and Polina Voronova. 2025. Mediating Art History Data Models for Native Linked Data Construction using ResearchSpace. In Digital Heritage, Stefano Campana, Daniele Ferdani, Holger Graf, Gabriele Guidi, Zackary Hegarty, Sofia Pescarin, and Fabio Remondino (Eds.). The Eurographics Association, Siena, Italy, 4 pages. https://doi.org/10.2312/dh.20253276
[3] Jochen Becker, H. V. D. Waal, L. D. Couprie, R. Fuchs, and E. Tholén. 1983. Iconclass: an iconographic classification system. Art Libraries Journal 8 (1983), 32–49. https://doi.org/10.1017/S0307472200003436
[4] Daniel J. Cohen. 2008. Creating Scholarly Tools and Resources for the Digital Ecosystem: Building Connections in the Zotero Project. First Monday 13, 8 (2008). http://www.uic.edu/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/2233/2017
[5] Li Ding, Joshua Shinavier, Zhenning Shangguan, and Deborah L. McGuinness. 2010. SameAs Networks and Beyond: Analyzing Deployment Status and Implications of owl: sameAs in Linked Data. In The Semantic Web - ISWC 2010 - 9th International Semantic Web Conference, ISWC 2010, Shanghai, China, November 7-11, 2010, Revised Selected Papers, Part I(Lecture Notes in Computer Science, Vol. 6496), Peter F. Patel-Schneider, Yue Pan, Pascal Hitzler, Peter Mika, Lei Zhang, Jeff Z. Pan, Ian Horrocks, and Birte Glimm (Eds.). Springer, Shanghai, China, 145–160. https://doi.org/10.1007/978-3-642-17746-010
[6] Martin Doerr. 2007. The CIDOC Conceptual Reference Model: A New Standard for Knowledge Sharing, In Challenges in Conceptual Modelling. Tutorials, posters, panels and industrial contributions at the 26th International Conference on Conceptual Modeling - ER 2007. Auckland, New Zealand, November 5-9, 2007. Proceedings. CIDOC CRM Documentation / Publication, 51–56. https://doi.org/10.13140/2.1.1420.6400
[7] Luciano Floridi. 2013. Information Quality. Philosophy and Technology 26 (2013), 1–6. https://doi.org/10.1007/s13347-013-0101-3
[8] Harry Halpin, Patrick J. Hayes, and Henry S. Thompson. 2015. When owl: sameAs isn't the Same Redux: Towards a Theory of Identity, Context, and Inference on the Semantic Web. In Modeling and Using Context - 9th International and Interdisciplinary Conference, CONTEXT 2015, Lanarca, Cyprus, November 2-6, 2015. Proceedings(Lecture Notes in Computer Science, Vol. 9405), Henning Christiansen, Isidora Stojanovic, and George A. Papadopoulos (Eds.). Springer, Larnaca, Cyprus, 47–60. https://doi.org/10.1007/978-3-319-25591-04
[9] Afraz Jaffri, Hugh Glaser, and Ian Millard. 2007. URI Identity Management for Semantic Web Data Integration and Linkage. In On the Move to Meaningful Internet Systems 2007: OTM 2007 Workshops, OTM Confederated International Workshops and Posters, AWeSOMe, CAMS, OTM Academy Doctoral Consortium, MONET, OnToContent, ORM, PerSys, PPN, RDDS, SSWS, and SWWS 2007, Vilamoura, Portugal, November 25-30, 2007, Proceedings, Part II(Lecture Notes in Computer Science, Vol. 4806), Robert Meersman, Zahir Tari, and Pilar Herrero (Eds.). Springer, Vilamoura, Portugal, 1125–1134. https://doi.org/10.1007/978-3-540-76890-640
[10] Maurizio Lenzerini. 2002. Data Integration: A Theoretical Perspective. In Proceedings of the Twenty-first ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems, June 3-5, Madison, Wisconsin, USA, Lucian Popa, Serge Abiteboul, and Phokion G. Kolaitis (Eds.). Association for Computing Machinery, 233–246. https://doi.org/10.1145/543613.543644
[11] Ruggero Longo and Elisabetta Scirocco. 2020. Sakralen Raum kartieren. Forschungsbericht der Max-Planck-Gesellschaft. https://www.mpg.de/16151367/biblhertzjb2020?c=152805.
[12] Ruggero Longo and Elisabetta Scirocco. 2025. Fragmented Objects / Fragmented Data. Quantitative methods and digital approaches in the project Mapping Sacred Spaces: forms, functions, and aesthetics in Medieval Southern Italy. Archeologia e Calcolatori 36, 2 (2025), 85–102. https://doi.org/10.19282/ac.36.2.2025.08
[13] Noemi Mauro et al. 2022. Exploring Semantically Interlaced Cultural Heritage Narratives. In Proceedings of the 33rd ACM Conference on Hypertext and Social Media (HT ’22). Association for Computing Machinery, 192–197. https://doi.org/10.1145/3511095.3536366
[14] Paul Mulholland, Peter van Kranenburg, Jason Carvalho, and Enrico Daga. 2024. Supporting the End-User Curation of Cultural Heritage Knowledge Graphs. In Proceedings of the 35th ACM Conference on Hypertext and Social Media, HT 2024, Poznan, Poland, September 10-13, 2024. Association for Computing Machinery, Poznan, Poland, 35–44. https://doi.org/10.1145/3648188.3675132
[15] Dominic Oldman and Diana Tanase. 2018. Reshaping the Knowledge Graph by Connecting Researchers, Data and Practices in ResearchSpace. In The Semantic Web - ISWC 2018 - 17th International Semantic Web Conference, Monterey, CA, USA, October 8-12, 2018, Proceedings, Part II(Lecture Notes in Computer Science, Vol. 11137), Denny Vrandecic, Kalina Bontcheva, Mari Carmen Suárez-Figueroa, Valentina Presutti, Irene Celino, Marta Sabou, and Elena Simperl (Eds.). Springer, Monterey, CA, USA, 325–340. https://doi.org/10.1007/978-3-030-00668-620
[16] Klaus Werner, Pietro Maria Liuzzo, and Alessandro Adamou. 2025. The Staccioli Digital Archive: Using Knowledge Graphs to power digital art history catalogues and art exhibitions. In Diversità, Equità e Inclusione: Sfide e Opportunità per l'Informatica Umanistica nell'Era dell'Intelligenza Artificiale. Proceedings of the XIV Annual Conference of AIUCD2025(Quaderni di Umanistica Digitale), Simone Rebora, Marco Rospocher, and Stefano Bazzaco (Eds.). AIUCD, Verona, Italy, 6 pages.
Do you like what you are reading? Subscribe to receive updates.
Unsubscribe anytime