II Networks of access and control

Computational ontologies for accessing, controlling, and disseminating knowledge in the cultural heritage sector: a case study

John Roberto Rodriguez

1. Introduction

Cultural heritage is rich in associations. Museum artworks contain semantically rich information that configures a semantic network: a collection of items has features and is related to other collections of items. For example, Sam Doyle was a self-taught African American artist. Doyle influenced Jean-Michel Basquiat, a leading member of the Neo-expressionist movement, who in turn collaborated with Andy Warhol. Basquiat once traded some of his own artworks to a gallery owner for a few of Doyle’s. Both Doyle and Basquiat were figurative artists. Another artist, Ed Ruscha, who has also cited Doyle’s influence, paid posthumous tribute to Doyle with his painting Where Are You Going, Man? (For Sam Doyle) in 1985. Doyle’s paintings and sculptures are kept at the American Folk Art Museum. This “semantic network” (see Figure 4.1) is not limited to a single collection but spans other related collections at different museums.

Nowadays, most of this information about art collections is often embedded within databases or within highly textual documents. This is a problem for researchers in digital humanities because it is difficult to extract, re-use, interpret, correlate, and compare the relevant information expressed implicitly in semi-structured and unstructured resources. Motivated by the above observation, researchers in Digital Humanities are working with computational ontologies to extract the implicit knowledge embedded in cultural resources and to make heterogeneous museum collections semantically interoperable. One of the most representative examples of this approach is the Europeana project.1 The heart of Europeana is to provide access to European cultural heritage by implementing the Europeana Data Model ontology.

Snippet of the semantic network on self-taught art

FIGURE 4.1 Snippet of the semantic network on self-taught art.

In general, an ontology is a form of knowledge conceptualization that makes this knowledge publicly available and reusable. From a technological point of view, a computational ontology is a collection of statements written in a formal language. Its purpose is to establish the relations between different concepts and specify logical rules for reasoning about them. Common components of ontologies include concepts (e.g. “Abstract Expressionism”), properties (e.g. “Abstract Expressionism is not focused on figures or imagery”), and relations (e.g. “Action Painting is an instance of the metaclass Abstract Expressionism”). Ontologies offer enhanced representation capabilities and they can also support reasoning operations that are at least partially similar to human reasoning. Through the use of a reasoner it is possible to derive new facts from existing ontologies. Reasoner is a software that works by inferring logical consequences from a set of explicitly asserted facts or axioms. Thus, ontologies allow us to make explicit domain assumptions: for example, “Abstract Expressionism has many stylistic similarities to the Russian artists of the early 20th century”. Ontologies formalize the intentional aspects of a domain in the form of a terminology or a T-box (e.g. an outsider artist can be defined as an artist with a mental illness), whereas the exten- sional part is formalized in the form of membership assertions or a В-box (e.g. Tarcisio Merati is an instance of the concept outsider artist).

In this chapter, we explore how computational ontologies make the process of accessing, controlling, and disseminating knowledge feasible in the cultural heritage sector. This chapter is divided into three major sections. The first section explains what Ontology Engineering is and how to extract ontologies from unstructured texts (Ontology Learning). Specifically, we analyse the extraction and population of ontologies by applying natural language analysis techniques to texts. The second section presents current research in ontology learning applied to the cultural heritage sector. We show concrete examples of available ontologies for museums, ontologies that have been developed for describing museum artefacts and objects, quantitative analyses of the art market using ontologies, web ontologies for modelling art collections, and new methods to provide personalized tour recommendations for museum visits. In the last section, we will take a case study to illustrate the utility of ontologies in preserving and disseminating cultural heritage.

2. Ontology engineering

The construction of ontologies by automatic or semiautomatic means is a popular area of research known as ontology engineering. Ontology engineering makes the process of learning ontologies feasible by establishing the most appropriate methodologies, methods, tools, and sources. It covers the conceptualization, design, implementation, and deployment processes.

When creating an ontology, it is important to follow ordered and defined steps, that is, to adopt a methodology. A methodology has to answer what, who, and when questions about performing activities in a development process (Gomez-Perez et al. 2007). In the field of ontology engineering, there are well- established methodologies to support the process of ontology development and maintenance: NeOn (Suarez-Figueroa et al. 2012), POEM (Ali & Khusro 2016), TERMINAE (Aussenac-Gilles et al. 2008), Termontography (Temmerman & Kerremans 2003), Bautista-Zambrana’s (2015) methodology, and An & Park’s (2018) methodology. All these methodologies include the following steps (Got- mare 2016):

  • • Identifying the purpose of the domain for which the ontology needs to be built.
  • • Capturing the concepts and the relationships between those concepts (ontology learning).
  • • Capturing the terms used to refer to those concepts and relationships (ontology population).
  • • Coding the ontology.

In the field of ontological engineering, ontology learning can be defined as the set of methods and techniques used for building an ontology in a semiautomatic or automatic fashion using several sources (Hazman et al. 2011). It is difficult to completely automate the process of ontology construction, but the technology should provide support for minimizing human effort. In order to make the process of inducing relevant ontological knowledge from data feasible, ontology learning uses methods from a diverse spectrum of fields such as machine learning, natural language processing (NLP), information retrieval, and text mining. Ontology learning methods, techniques, and approaches, therefore, serve the purpose of supporting the ontology engineer in the task of creating and maintaining an ontology (Cimiano et al. 2009).2

Most ontology learning approaches follow a model named the Ontology Learning Layer Cake (OLC). The OLC model aims to learn ontologies by using sequential steps or tasks because it was inspired by the “divide and conquer” algorithm design paradigm. These tasks are organized according to their increasing complexity within the process of ontology learning. Broadly speaking, there are three main tasks in ontology learning:

  • • Term extraction: The identification of terms that are relevant to the domain.
  • • Relation extraction: Once the relevant terms are extracted, interactions between them must then be established.
  • • Axioms extraction: Axioms are propositions that are always taken as true with respect to a specific domain.

To obtain knowledge of the domain covered by texts, ontology learning often employs NLP techniques. There are three different approaches to ontology learning from text based on the technique used in addition to NLP. First, the statistical approach, which uses text-mining techniques to discover concepts and taxonomical relations in documents. Second is the pure NLP approach, which uses shallow or deep NLP techniques to discover the relations between syntactic entities. Shallow NLP techniques generate a partial analysis of texts by including chunking and part-of-speech tagging. Deep NLP techniques generate a full analysis of texts by including a syntactic parser and very fine-grained aspects of the language such as coreference and anaphora resolution. Third is the integrated approach, which merges both statistical and NLP approaches.

The quality of the input is one of the parameters to be taken into account when devising an ontology. This is particularly important for ontology learning from text because NLP techniques depend on corpus quality, as has been shown by many previous studies (e.g. Eckart et al. 2012). If the text is not well-formed, NLP tasks such as syntactic parsing and anaphora resolution are not feasible. For this reason, the source text used for ontology learning should be well-formed, that is, a body of scientific books, papers, magazines, or web pages. In the latter case, it is important to build additional corpora from the web through scraping and crawling techniques. Web scraping is the process of iteratively gathering pages from the web in an automated fashion. Web crawling attempts to download only those pages that are on the domain of the target ontology. Crawlers rely on the fact that pages about a topic tend to contain links to other pages on the same topic (Liu et al. 2015).

Some ontology learning approaches do not derive schematic structures, but derive facts from text by using ontology population methods. Ontology population is a crucial part of knowledge base construction and maintenance. It consists of identifying the key terms in the text and then relating them to concepts in the ontology. Typically, this task is carried out from a linguistic approach. Linguistic-based approaches aim to obtain knowledge from the domain by applying NLP techniques and tools to text analysis. Typical NLP techniques include the linguistic pre-processing of texts (e.g. tokenization, part-of-speech tagging, lemmatization, stemming, sentence splitting, and morphological analysis), the identification of named entities in texts (e.g. persons, organizations, locations, and date/time expressions), automatic term recognition, the analysis of sentences to derive their syntactic structure (parsing and chunking), and the extraction of lexico-syntactic patterns from texts to identify relationships between terms and entities.

A number of well-designed, well-defined, and compatible formal resources (languages and tools) are used to represent and share ontologies. For example, Resource Description Framework (RDF), Web Ontology Language (OWL), and SPARQL are three of the standard core languages for the Semantic Web. Firstly, the RDF is a language for representing binary relations between two resources in the web. The two resources (subject and object) and the relation (predicate) form a triple: for example, Picasso —> was-born-in —> Spain. Secondly, OWL/OWL2 is a language for making ontological statements whose syntax and formal semantics are derived from description logics. Thirdly, a number of query languages have been developed to extract information from RDF and OWL, including SPARQL for RDF and SQWRL for OWL.

3. Ontologies for cultural heritage

Ontologies have found fertile ground in the cultural heritage domain due to the need to preserve, conserve, curate, and disseminate physical and digital objects. Most of the resources in the heritage sector are huge and heterogeneous. Cultural heritage includes “highly structured, very unstructured, and semi-structured data or information obtained from both authorized and unauthorized sources and involving multimedia data including text, audio, and video data” (Bing et al. 2014). The sector is characterized by a complex data integration problem for which a solution may be found through the development of standard relational schema for museums in the hands of international organizations such as the International Council of Museums (ICOM), the International Federation of Library Associations (IFLA), the International Council of Archives (ICA), and the Dublin Core Consortium. Additionally, there are targeted initiatives to ontologize selected parts of this domain. This section describes both reference ontology models and four typical applications of ontologies: the preservation and dissemination of cultural heritage, the development of high-level software tools for cultural heritage, the establishment of smart museums, and the description of art market data.

3.1. Core reference ontologies for cultural heritage

There are several well-known attempts to provide mechanisms able to perform knowledge representation across many cultural heritage sources, such as Conceptual Reference Model (CRM), Visual Resource Association Core (VRA), and European Data Model (EDM). The CRM is “a formal ontology intended to facilitate the integration, mediation and interchange of heterogeneous cultural heritage information” (Le Boeuf et al. 2018). It was developed by interdisciplinary teams of experts under the aegis of the International Committee for Documentation (CIDOC) of the ICOM. CIDOC CRM provides a set of reference standards used for describing artefacts in museum collections and other cultural institutions. The CIDOC CRM was published in 2006 as an ISO standard (ISO 21127:2006). ISO 21127 is the reference ontology standard for the interchange of cultural heritage information and is at the root of the CRM model. Because of this, CIDOC CRM provides a common and extensible ontology that any cultural heritage information can be mapped onto.

The VRA is a data standard for the description of metadata about cultural works and their images. To enable linked data capable VRA metadata, a VRA ontology was designed by the VRA Oversight Committee. Additionally, vocabulary from a number of different ontologies was mapped onto the VRA ontology to support increased interoperability, extensible Markup Language (XML) schema was used as a basis for developing the VRA ontology as a means of making it highly interoperable across the web. According to Mixter (2014), the existing data can be successfully converted into RDF by using an Extensible Stylesheet Language Transformations (XSLT) style sheet.

The Europeana Data Model (Europeana 2013) is an ontology-based framework that is suitable for the description of cultural objects. It is available as Linked Open Data and can be explored and queried through the SPARQL application programming interface (API). This ontology was developed within the framework of Europeana and is a multilingual research initiative for facilitating user access to integrated European cultural and scientific heritage content. Europeana aims to integrate data from different sources in a single data set and provide users with a unified view of the data. The metadata for all the objects in the Europeana portal are open and can be freely downloaded via the API.

Liu et al. (2017) evaluated those three popular cultural heritage ontologies with respect to their abilities to represent works of art. They found that the CIDOC CRM ontology is able to capture a wide range of concepts in many different ways, the EDM ontology is more appropriate for creating and aggregating simpler semantic descriptions, and the VRA Core ontology is the most appropriate for cataloguing visual resources.

3.2. Ontologies for the preservation and dissemination of cultural heritage

Ontologies are often employed for the preservation and dissemination of cultural heritage. OPPRA, Ontology of Paintings and Preservation of Art (Odat 2014), was developed to support the information integration and analysis requirements of art conservators. It provides a common machine-readable formal representation of knowledge in the domain of art and paint preservation. The PARCOURS project (Niang et al. 2017) is an application ontology whose objective is to provide a common reference point that would facilitate information sharing among different conservation—restoration professionals. It is capable of offering a consensual framework for the formal representation of conservation-restoration data. The Conservation Reasoning (CORE) ontology (Moraitou & Kavakli 2018) integrates in different information to form the body of knowledge relevant to taking thoughtful decisions on the treatment and care of cultural heritage objects. In Moraitou et al. (2018), the CORE ontology and the Semantic Sensor Network (SSN) ontology are integrated to create a new merged ontology that combines conservation procedures, data, and rules with sensor and environmental information.

A number of ontologies have been developed for the organization and dissemination of intangible cultural heritage (ICH). Dou et al. (2018) constructed an ICH ontology to normalize a portion of the ICH of China. Domain knowledge was extracted from ICH text data using NLP technology and ICH knowledge was presented based on the ICH knowledge graph. The domain ontology of Tujia brocade (Zhao et al. 2017) provides a good platform for promoting the sharing and spreading of knowledge about Tujia brocade, an example of China’s national ICH. The Knowledge Discovery Multi-agent System (KDMAS) design methodology was developed by Kadar & Tulbure (2017) to automatically populate ontologies. They built a framework consisting of two modules. The first module downloads documents from the web using a thesaurus as a reference when searching for documents. The second module uses a defined gazetteer to extract the desired information, which is then added as instances to a customized web-based ontology created with Protege. The domain used to evaluate the KDMAS framework was Cultural Heritage, an ontology that was created for the semantic description of ancient socketed axes.

Several other ontologies have been developed for the preservation of digital resources. In the scope of the PERICLES project, Lagos et al. (2015) focused on the long-term digital preservation of digital artworks and experimental scientific data by capturing digital ecosystems, that is, the environments in which digital objects are created, managed, and used. The Linked Resource Model (LRM) ontology was designed as a principled way to model digital ecosystems. LRM defines the conditions required for a change to happen in the ecosystem and/or the impact of such a change. For example, ontological dependencies describe the context under which change in one entity has an impact on other entities in the ecosystem. Tibaut et al. (2018) investigated the development of an ontology for the domain of heritage buildings considering different methodologies and different input knowledge. Their research implemented four of the six development steps that characterize the ontology learning process under the Methontology approach: (1) specification of the purpose of the ontology; (2) the collection of available knowledge; (3) a conceptualization phase; and (4) the reuse of existing ontologies.

3.3. Ontologies for software development

Cultural heritage ontologies are also employed to support the development of high-level software tools. The richness and diversity of cultural heritage objects have motivated the use of ontologies for efficient indexing and searching in information retrieval systems. For example, the Big Ancient Mediterranean (see Horne, this volume) project provides an extensible digital framework that uses semantic web technologies for connecting textual, geospatial, and network data sources from the ancient world. Kambau & Hasibuan (2017) built two ontologies as the first step in the development of the Concept-based Multimedia Information Retrieval System (MIRS) for Indonesia’s cultural heritage. Garozzo et al. (2017) designed a Cultural heritage Tool (CulTO) to support the curation of photographic data and text documents on historical buildings. CulTO relied on a computational ontology to successfully identify and characterize historical religious buildings. The CulTO computational ontology was designed using standard ontologies and schemas (e.g. CIDOC CRM).

In recent years, Augmented Reality (AR) has received much attention as an interactive medium for requesting and accessing information in the cultural heritage domain. Kim et al. (2017) developed a mobile AR application based on ontologies to provide contextual information about cultural heritage sites, such as the person who created the site and the events that occurred at that location. A cultural heritage ontology aggregates heterogeneous data by using the Korea Cultural Heritage Data Model. In the same vein, the TAIS Tourist Assistant (Smirnov et al. 2017) is a system that recommends cultural heritage to tourists based on their preferences and location. The system consists of a set of services joined together by an RDF ontology that defines the main concepts and relationships for the interaction of the TAIS components. Information about cultural heritage is extracted from different Internet sources such as Wikipedia, Wikivoyage, and Panoramic. Lin & Lin (2017) developed CLIPS, an ontology- based expert system to encode relationships between artworks and artists. The embedded ontology was composed of information provided by the Asia University Museum of Modern Art, such as art types, subclasses of art types, art periods, prestigious exhibitions, awards for artworks, the importance of artworks, and artists’ life stories.

Natural Language Generation Systems (NLGs) applied to the Cultural Heritage domain were investigated in Piccialli et al. (2017). Piccialli et al. propose Natural OWL, an NLG engine able to automatically build structured textual descriptions of cultural objects based on art domain ontologies. The ontologies include assertional knowledge such as a domain template and domain instances. The ontology output was represented by lists of relevant terms and sentences. The Natural OWL engine and its ontologies were integrated in the FEDRO platform to support a cultural exhibition of “talking” sculptures held in Southern Italy. Capodieci et al. (2016) introduced the Semantic Enterprise Service Bus (ESB), a software system that can extract semantically related content from heterogeneous data sources. The prototype presented was designed to be used in the field of cultural heritage by integrating the EDM ontology. In a broad sense, ESB works by finding matches between the metadata in two databases related to ancient coins and inscriptions and the (sub)properties and classes of EDM.

Finally, ontologies can be transformed automatically into other kinds of resources. For example, Hnatkowska & Woroniecki (2018) presented a method for transforming ontologies provided in RDF and OWL/OWL2 into runnable object-oriented code (Groovy). According to its authors, “the fact that Groovy is a target transformation language does not limit the usage of the approach to applications written only in Groovy” (p. 281).

3.4. Ontologies for the creation of smart museums

Cultural heritage ontologies are a common component of smart museums. There are currently pilot initiatives at the Smart Museum of Art at the University of Chicago, China’s National Smart Museum, the National Palace Museum of Korea, as well as at several museums in New York.

Petrina et al. (2017) developed an ontological model for the structural description of the objects collected at the Smart Museum of Everyday Life History at Petrozavodsk State University in Russia. The proposed ontological model was based on CIDOC CRM for the representation of historical information about the exhibits. The ontology provides structural rules for creating the required semantic network, which was represented using the RDF. The researchers’ ultimate goal was to develop cultural heritage services based on an ontology modelling that personalizes access to the museum collection. Hajmoosaei & Skoric (2016) introduced an ontology development methodology to enable an intelligent search for heritage entities. Based on this methodology, they created two ontologies for a New Zealand museum. The first ontology is a top-level ontology that builds on Europeana and CRM models by adding unique subclasses and properties. The second ontology is a local ontology designed to capture the semantics of New Zealand soldiers who took part in the First and Second World Wars.

3.5 Ontologies for art market research

Ontologies can be used to describe art market data in art market research. Fili- piak et al. (2016) presented a method for quantitative art market research using ontologies. They proposed mixing a standard econometric analysis with the use of an art market ontology in order to build precise art market indices. Their approach uses features extracted from auction house websites, DBpedia, Europeana, the WikiArt Visual Art Encyclopedia, and computer vision tools that automatically predict the style of a painting.3 The resulting art market ontology concentrates on the description of paintings that appeared in auctions and the central entity is therefore the Lot.

4. Ontologizing outsider art

In this section, we will take a case study to illustrate the utility of ontologies in preserving and disseminating cultural heritage. Specifically, we will describe the use of ontologies for the standardization of knowledge of outsider art.

The history of art can be divided into two broad categories: mainstream art and outsider art. Compare Pablo Picasso’s Self-Portrait, 1972 (Fuji Television Gallery, Tokyo) to an example of outsider art in Figure 4.2. Mainstream art, also called traditional art, is part of the dominant culture. Perhaps this is the reason why, until quite recently, only mainstream art was exhibited at galleries and museums. In contrast, outsider art has always been the “other art”, or the art made by people who are not regarded as part of the mainstream art world. Today, as well as including artists with disabilities or mental illness, the term is increasingly applied to “the others”: those on the margins of art and society. As a consequence, prisons, asylums, streets, refugee camps, and boarding schools seem to be the locus for outsider art. In general, this issue can be summarized in the dichotomy between concepts such as high and low art, inside and outside art, rational and irrational art, thoughtful and naive art, rule-based and anarchic art, tutored and untutored artists, trained and untrained artists, and schooled and self-taught creators.

For a large part of the mainstream, outsider art is unsightly rubbish. Flowever, paradoxically, mainstream art is interested in acquiring examples of outsider art. For example, American vanguard artists found inspiration in the work of their marginalized peers. The exhibition Outliers and American Vanguard Art, at the National Gallery of Art in January 2018, shows how the dialogue between vanguard and self-taught artists has been defined by contradiction. In the same vein, one of the topics discussed at the outsider art fair held in New York in 2014 was Basquiat as Self-Taught Artist. Jean-Michel Basquiat, one of the most widely known contemporary artists, led a life of struggle and died of a heroin overdose in his art studio. Like outsider artists, Basquiat’s work is characterized by the use of found materials, the repetition of symbols, the use of text in his drawing, and his interest in dichotomies: wealth versus poverty, integration versus segregation, and so on. Basquiat, a reference for the mainstream, collected and was inspired by the works of Sam Doyle, a self-taught artist from South Carolina. Ironically, while Basquiat’s work sells for millions, Doyle’s work does not.

Luckily, outsider art is being recognized as art these days and its visibility has increased dramatically in recent years. In 2013, the new Whitney Museum of American Art’s inaugural exhibition included a work by Bill Traylor. In 2014, the Metropolitan Museum of Art in New York included in its catalogue 57 works by outsider artists. Christie’s and Sotheby’s, the Western world’s two largest auction houses, are holding auctions of works by outsider artist for thousands, or even millions of dollars. In 2015, Dubuffet’s Paris Polka (1961) was sold for EUR 21.8 million. Boxer, a sculpture by William Edmondson, has an estimated value of EUR 220,000. But Christie’s and Sotheby’s are providing collectors with an opportunity to snap up more accessible works too, often from as little as EUR 1,000.

Outsider artist

FIGURE 4.2 Outsider artist: Self-portrait by Daniel Saraclio, 2018.

Outsider art is undeniably innovative. Joaquim Corral, director of the mAB-Art Brut Museum in Barcelona, said recently that innovation in the Fine Arts starts from outsider art. This understanding is shared by Patriotta & Hirsch (2016), for whom art innovation stems from mainstreamers, mavericks, conventional novices, and outsiders, all working together as a network. But outsider art still remains a nebulous domain and a deeply problematic notion. In many respects, the world of outsider art is a hermetic part of the world of creativity and, by extension, of society. The very idiosyncrasy of outsider art has made it difficult to classify and provide a clear appellation for the genre. Over the years, the term has been used to describe artwork produced by psychiatric patients, spiritualists, the homeless, self-taught visionaries, illiterate persons, eccentrics, recluses, criminals, and others beyond the perceived margins of society. Today a plethora of deceptive terms are used to describe it: art brut, art of madmen, art singulier, autistic art, contemporary folk art, faux naive art, folk art, fresh invention, grass-roots art, intuitive art, marginal art, mediumistic art, naive art, neuve invention, non-traditional folk art, primitive art, primitivism, pseudo-naive art, psychopathologic art, psychotic art, raw art, self-taught art, vernacular art and visionary art. For example, the term neuve invention serves as a category to include creators whose work seemed difficult to label as outsider art; the term marginal art refers to the work of contemporary artists that lie between outsider art and mainstream art; and naive art, also spelled naif art, refers to art created by untrained artists.

As a result, outsider art remains the subject of highly diverse debates as to its meaning. There are those who believe that outsider art is a tenable concept. For example, Marcus Davies (2007) states that the use of the term is here to stay. Fie says that “viewed in contrast to long-established assumptions regarding the role of art, artists, and the transmission of dominant values, outsider art may be understood as a genre born of negation”. Thomas Roske (2005) does not reject the concept, but questions the attitude of seeing outsider art as the only true art. Chris Wiley (2013) sets out four subcategories “under the umbrella of this broader rubric”: traditional folk practices, the mentally ill, idiosyncratic art, and metaphysic art (Wiley 2013). On the other side of the issue, there is a small group of skeptics who question the authenticity of the concept. This means that they hold that that none of the objects exhibited under that label are artistic vehicles of artworks (Davies 2009). James Elkins (2006) goes far beyond that by claiming that the term is an oxymoron and, consequently, there is no such thing as outsider art: “the first thing that needs to be said about outsider art is that it does not exist”. According to the art critic Katherine Jentleson (2014), the distinctions between outsiders and insiders appear “increasingly dubious”.

This leads us to conclude that there is a need to perform an explicit terminological standardization of the outsider art domain. This standardization process is closely related to the ontologization process. Zouaq (2011) redefines ontologization as “the ability to build a bridge between the language level and the abstract/conceptual level” or, in other words, “the ability to filter important concepts and relationships among the mass of extracted knowledge”. Meyer & Gurevych (2012) define ontologization as the formation of ontological concepts and relationships from harvested knowledge. Zouaq et al.

(2011) define the ontologization task as the ability to build a bridge between the natural language level and the abstract/conceptual level. Similarly, for Pri- yatna & Villazon-Terrazas (2012), ontologization is a re-engineering procedure that transforms non-ontological resource terms into ontology representational primitives (e.g. the transformation of relational databases, thesauri or lexicons into RDF). Boualem et al. (2017) differentiate between conceptualization and ontologization: conceptualization is a description of concepts and relationships, and ontologization involves expressing constraints (axioms) in first-order logic. Pantel & Pennacchiotti (2008) differentiate between extracting binary semantic relations from textual resources (harvesting) and linking those relations into a semantic repository upon which knowledge representation reasoning engines can operate (ontologizing). Finally, Tibaut et al. (2018) consider the ontologization of cultural data to be a subdomain of cultural heritage by which a team of experts promotes cultural standardization and institutionalization: “Concerning the process of ontologization, the team of experts applied strict principles to admit only concepts that serve the functionality of global information integration, and other, more philosophical restrictions about the kind of discourse to be supported” (p. 150).

Ontologies can play a significant role in the terminology standardization of outsider art. If terminology is the activity of collecting, describing, and organizing terms in a resource, ontologies provide a type of formal conceptual structure that can result in more controlled, consistent, logical, and systematic terminological resources (Bautista-Zambrana 2015). One of the main advantages of using ontologies for the standardization of outsider art terminology is their explicitness, that is, the ability to represent knowledge in a clear and comprehensive way. Ontologies make the relationships between concepts explicit, which is a necessary task for the sharing and normalization of knowledge about outsider art. That means that in order to understand terms it is essential to know how terms are related to one another. For example, the term “mysticism” can help to differentiate “outsider art” from “visionary art” within the domain.

An ontology for outsider art will allow for the specification of the set of concepts and the relationships between those concepts that are possible or permissible in the domain. Therefore, expressing outsider art knowledge as an ontology is an appropriate approach to making domain knowledge understandable and distinguishable from similar domains. Additionally, the outsider art ontology will be used to organize and express domain knowledge in a machine-readable format. It will be suitable for discovering implicit facts, relations, and contradictions by using reasoning engines. Encoding knowledge about outsider art can increase readability, allowing computers to quickly identify the relationships between different concepts. In general, the resulting ontology will contribute to determining the semantic boundaries of outsider art. Through the standardization of specialized concepts and their designations across the domain, it will be possible to generate new knowledge that allows for the evolution of outsider art in relation to the mainstream.


  • 1 www.europeana.eu
  • 2 An approach is a theory, a method is an application of an approach, and a technique is a single activity that comes from a procedure.
  • 3 www.wikiart.org/


Ali, S. and Khusro, S., (2016). POEM: Practical ontology engineering model for semantic web ontologies. Cogent Engineering [online]. 3(1). https://doi.Org/10.1080/23311916.2 016.1193959

An, J. and Park, Y., (2018). Methodology for automatic ontology generation using database schema information. Mobile Information Systems [online], http://doi.org/10.1155/2018/ 1359174

Aussenac-Gilles, N.. Despres, S. and Szulman, S., (2008). The TERMINAE method and platform for ontology engineering from texts. In: P. Buitelaar and P. Cimiano, eds. Ontology learning and population: Bridging the gap between text and knowledge. Amsterdam: IOS Press, pp. 199-223.

Bautista-Zambrana, M., (2015). Methodologies to build ontologies for terminological purposes. Procedia—Social and Behavioral Sciences. 173, 264-269.

Bing, L., Chan, К. C. and Carr, L., (2014). Using aligned ontology model to convert cultural heritage resources into semantic web. 2014 IEEE International Conference on Semantic Computing [online]. Newport Beach, CA. pp. 120-123. doi: 10.1109/ ICSC.2014.39

Boualem, S. A., Meryem, B. and Fatima, N., (2017). Maintenance & information security ontology. 2017 4th International Conference on Control, Decision and Information Technologies (CoDIT) [online], Barcelona, pp. 312-317. doi: 10.1109/CoDIT.2017.8102610

Capodieci, A., Mainetti, L. and Carrozzo, S., (2016). Semantic enterprise service bus for cultural heritage. In: 12th International Conference on Innovations in Information Technology (IIT) [online], Al-Ain, UAE. pp. 156-163. doi: 10.1109/INNOVAT1ONS. 2016.7880044

Cimiano, P., Maedche, A., Staab, S. and Volker, J., (2009). Ontology learning. In: S. Staab and R. Studer, eds. Handbook on ontologies. Berlin: Springer, pp. 245-267.

Davies, D., (2009). On the very idea of‘outsider art’. British Journal of Aesthetics. 49(1), 25-41.

Davies, M., (2007). On outsider art and the margins of the mainstream [online]. 6 April. Available from: www.ibiblio.org/frenchart/

Don, J., Qin, J., Jin, Z. and Zhuang, L., (2018). Knowledge graph based on domain ontology and natural language processing technology for Chinese intangible cultural heritage. Journal of Visual Languages & Computing. 48, 19-28.

Eckart, T., Quasthoff, U. and Goldhahn, D., (2012). The influence of corpus quality on statistical measurements on language resources. In: N. Calzolari, K. Choukri, T. Declerck et al., eds. Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC’12) [online]. Istanbul, Turkey, pp. 2318-2321. Available from: www.lrec-onf.org/proceedings/lrec2012/pdf/476_Paper.pdf

Elkins, J., (2006). NaYfs, faux-nai'fs, faux-faux-naifs, would-be faux-nai'fs: There is no such thing as outsider art. In: [. Thompson, ed. Mundos interiores al descubierto/inner worlds outside. Madrid: Fundacio Caixa de Pensions and Dublin: Irish Museum of Modern Art. pp. 71-78.

Europeana., (2013). Europeans data modeI primer [online|. Isaac, A., ed. Available from: https://pro.europeana.eu/files/Europeana_Professional/Share_your_data/Technical_ requirements/EDM_Documentation/EDM_Primer_130714.pdf

Filipiak, D., Agt-Rickauer, H., Hentschel, C., Filipowska, A. and Sack, H., (2016). Quantitative analysis of art market using ontologies, named entity recognition and machine learning: A case study. In: W. Abramowicz, R. Alt and B. Franczyk, eds. Business Information Systems: 19th International Conference, BIS 2016, Leipzig, Germany, 6-8 July 2016, Proceedings [online], pp. 79-90. https://doi.org/10.1007/978-3-319-39426-8_7

Garozzo, R., Murabito, F., Santagati, C., Pino, C. and Spampinato, C., (2017). CulTO: An ontology-based annotation tool for data curation in cultural heritage. ISPRS International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences [online]. 42(2), 267-274. https://doi.org/10.5194/isprs-archives-XLII-2-W5-267-2017

Gomez-Perez, A., Fernandez-Lopez, M. and Corcho, O., (2007). Ontological engineering: With examples from the areas of knowledge management, e-commerce and the semantic web. New York: Springer.

Gotmare, P., (2016). Methodology for semi-automatic ontology construction using ontology learning: A survey. International Journal of Computer Applications. 2, 1-3.

Hajmoosaei, A. and Skoric, P., (2016). Museum ontology-based metadata. In: 2016 IEEE Tenth International Conference on Semantic Computing (ICSC), 4-6 February 2016, Laguna Hills, CA, USA [online], pp. 100-103. doi: 10.1109/ICSC.2016.74

Hazman, M., El-Beltagy, S. and Rafea, A., (2011). A survey of ontology learning approaches. International Journal of Computer Applications [online]. 22(9), 36-43. doi: 10.5120/2610-3642

Hnatkowska, B. and Woroniecki, P., (2018). Transformation of OWL2 property axioms to Groovy. In: A. Tjoa, L. Bellatreche, S. Biffl, J. van Leeuwen and J. Wiedermann, eds. SOFSEM 2018: Theory and Practice of Computer Science: 44th International Conference on Current Trends in Theory and Practice of Computer Science, Krems, Austria, January 29—February 2 2018, Proceedings. Cham: Springer, pp. 269-282.

Jentleson, K., (2014). Cracks in the consensus: Outsider artists and art world ruptures. In: T. J. Lax et al., eds. When the stars begin to fall: Imagination and the American South. New York: The Studio Museum in Harlem.

Kadar, M. and Tulbure, A., (2017). An automated knowledge discovery framework with multi-agent systems-KDMAS. International Conference on Engineering, Technology and Innovation, ICE/1TMC 2017, Funchal, Portugal, 21-29 June 2011 [online], pp. 1262- 1267. doi: 10.1109/ICE.2017.8280025

Kambau, R. A. and Hasibuan, Z. A.. (2017). Concept-based multimedia information retrieval system using ontology search in cultural heritage. Second International Conference on Informatics and Computing (ICIC), 1-3 November 2011, Jayapura-Papua, Indonesia [online],doi: 10.1109/IAC.2017.8280623

Kim, H., Matuszka, T, Kim, J., Kim, ). and Woo, W., (2017). Ontology-based mobile augmented reality in cultural heritage sites: Information modeling and user study. Multimedia Tools and Applications. 76(8), 26001—26029.

Lagos, N., Waddington, S. and Vion-Dury, J.-Y., (2015). On the preservation of evolving digital content: The continuum approach and relevant metadata models. In: E. Garou- fallou, R. Hartley and P. Gaitanou, eds. Metadata and Semantics Research: Communications in Computer and Information Science: 10th International Conference, MTSR 2016, 22-25 November 2016, Gottingen, Germany, Proceedings [online]. Cham: Springer, pp. 15—26.

Le Boeuf, P., Doerr, M., Ore, C. and Stead, S., (2018). Definition of the CIDOC conceptual reference model. ICOM/CIDOC Documentation Standards Group, Continued by the

CIDOC CRM Special Interest Group. Version 6.2.4. November [online]. Available from: www.cidoc-crm.org/sites/default/files/2018-10-26%23CIDOC%20CRM_


Lin, F. and Lin, Y., (2017). An ontology-based expert system for representing cultural meanings: An example of an art museum. 2017 Pacific Neighborhood Consortium Annual Conference and Joint Meetings (PNC), 7-9November 2017, Tainan, Taiwan [online], pp. 116—121. doi: 10.23919/PNC. 2017.8203531

Liu, D., Bikakis, A. and Vlachidis, A., (2017). Evaluation of semantic web ontologies for modelling art collections. New Trends in Databases and Information Systems: ADBIS 2017: Short Papers atid Workshops, AMSD, BigNovelTI, DAS, SW4CH, DC, Nicosia, Cyprus, 24-27 September 2017, Proceedings [online], pp. 343-352.

Liu, L., Peng, T. and Zuo, W., (2015). Topical web crawling for domain-specific resource discovery enhanced by selectively using link-context. International Arab Journal of Information Technology. 12(2), 196-204.

Meyer, С. M. and Gurevych, L, (2012). OntoWiktionary: Constructing an ontology from the collaborative online dictionary wiktionary. In: M. T. Pazienza and A. Stellato, eds. Semi-automatic ontology development: Processes and resources. Hershev, PA: IGI Global, pp. 131-161.

Mixter, J., (2014). Using a common model: Mapping VRA Core 4.0 into an RDF ontology. Journal of Library Metadata. 14(1), 1-23.

Moraitou, E., Aliprantis, J. and Caridakis, G., (2018). Semantic preventive conservation of cultural heritage collections. In: A. Bikakis, S. Jean, B. Markhoff and A. Mosca, eds. Semantic Web for Cultural Heritage 2018 (SW4CH 2018): Proceedings of the 3rd International Workshop on Semantic Web for Cultural Heritage, 3 June 2018, Heraklion, Crete [online]. Available from: http://ceur-ws.org/Vol-2094/

Moraitou, E. and Kavakli, E., (2018). Knowledge management using ontology on the domain of artworks conservation. In: M. Ioannides, ed. Digital Cultural Heritage: Final Conference of the Marie Sktodowska-Curie initial Training Network for Digital Cultural Heritage, ITN-DCH 2017, Olimje, Slovenia, 23-25 May 2017, Revised Selected Papers [online], Cham: Springer, pp. 50-62.

Niang, C., Marinica, C., Markhoff, B. et al., (2017). Supporting semantic interoperability in conservation-restoration domain: The PARCOURS project. Journal on Computing and Cultural Heritage [online]. 10(3). http://doi.org/10.1145/3097571

Odat, S., (2014). A semantic e-science platform for 20th century paint conservation. Ph.D. thesis, the University of Queensland, https://doi.org/10.14264/uql.2014.426

Pantel, P. and Pennacchiotti, M., (2008). Automatically harvesting and ontologizing semantic relations. In: P. Buitelaar and P. Cimiano, eds. Ontology learning and population: Bridging the gap between text and knowledge. Amsterdam: IOS Press, pp. 171-195.

Patriotta, G. and Hirsch, P., (2016). Mainstreaming innovation in art worlds: Cooperative links, conventions and amphibious artists. Organization Studies. 37(6), 867-887.

Petrina, O., Volokhova, V., Yalovitsyna, S., Varfolomeyev, A. and Korzun, D., (2017). On semantic network design for a smart museum of everyday life history. Proceedings of the 20th Conference of Open Innovations Association FRUCT, 3-7 April 2017, St. Petersburg, Russia [online], pp. 676-680.

Piccialli, F., Marulli, F. and Chianese, A., (2017). A novel approach for automatic text analysis and generation for the cultural heritage domain. Multimedia Tools and Applications. 76(8), 10389-10406.

Priyatna, F. and Villazon-Terrazas, B., (2012). Building ontologies by using re-engineering patterns and R2RML mappings. Proceedings of the 3rd International Conference on Ontology

Patterns (WOP2012), 12 November 2012, Boston, US/f [online], pp. 109-120. Available from: http://ceur-ws.org/Vol-929/

Roske, T., (2005). The only true art? Outsider art as an opposition. International Conference on History and Future of Outsider Art. Helsinki: Museum of Contemporary Art Kiasma. pp. 56-59. Available from: http://archiv.ub.uni-heidelberg.de/artdok/5063/!/ Roeske_The_only_true_art_2005.pdf

Smirnov, A., Kashevnik, A. and Ponomarev, A., (2017). Context-based infomobility system for cultural heritage recommendation: Tourist assistant-TAIS. Personal Ubiquitous Computing. 21(2), 297-311. https://doi.org/10.1007/s00779-016-0990-0

Suarez-Figueroa, M. C., Gomez-Perez, A. and Fernandez-Lopez, M., (2012). The NeOn methodology for ontology engineering. In: M. C. Suarez-Figueroa, A. Gomez- Perez, E. Motta and A. Gangenti, eds. Ontology engineering in a networked world. Berlin: Springer, pp. 9-34.

Temmerman, R. and Kerrentans, K., (2003). Termontography: Ontology building and the sociocognitive approach to terminology description. In: E. Hajicova, A. Kotesovcova and J. Mlrovsky, eds. Proceedings 17th International Congress of Linguists, Prague, 24-29 July 2003. Prague: Matfyzpress. pp. 1-10.

Tibaut, A., Kaucic, B., Dvornik Perhavec, D., Tiano, P. and Martins,J., (2018). Ontologizing the heritage building domain. In: M. Ioannides, J. Martins, R. Zarnic and V. Lim, eds. Advances in Digital Cultural Heritage: International Workshop, 28 June 2017, Funchal, Madeira, Portugal, Revised Selected Papers [online]. Cham: Springer, pp. 141-161.

Wiley, C., (2013). Outside in. Mousse Magazine [online]. 40. Available from: http://mous- semagazine.it/chris-wiley-outside-in-2013/

Zhao, G., Luo, Z., He, H., Li, Y., Xia, ). and Zan, H., (2017). Research and realization of ontology-based Tujia brocade knowledge base system. 3rd IEEE International Conference on Computer and Communications (ICCC), 13-16 December 2017, Chengdu, China [online], pp. 2569-2573. doi: 10.1109/CompComm.2017.8322999

Zouaq, A., (2011). An overview of shallow and deep natural language processing for ontology learning. In: W. Wong, W. Liu and M. Bennantoun, eds. Ontology learning and knowledge discovery using the web: Challenges and recent advances. Hershey, PA: IGI Global, pp. 16-37.

Zouaq, A., Gasevic, D. and Hatala, M., (2011). Ontologizing concept maps usinggraph theory. Proceedings of the 2011 ACM Symposium on Applied Computing (SAC), 21-24 March 2011, TaiChung, Taiwan [online], pp. 1687—1692. https://doi.org/10.1145/1982185.1982537

< Prev   CONTENTS   Source   Next >