Time filter

Source Type

Paris, France

Mondeca is a leading provider of software based on semantic technologies. It has been recognized by Gartner as a top vendor in the Semantic Web as well as in the Taxonomy and Ontology Management categories. Wikipedia.

Battistelli D.,Paris-Sorbonne University | Cori M.,French National Center for Scientific Research | Minel J.-L.,French National Center for Scientific Research | Teissedre C.,French National Center for Scientific Research | Teissedre C.,Mondeca
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

Unlike most approaches in the field of temporal expressions annotation, we consider that temporal adverbials could be relevant units from the point of view of Information Retrieval. We present here the main principles of our semantic modeling approach to temporal adverbial units. It comprises two steps: functional modeling (using a small number of basic operators) and referential modeling (using calendar intervals). In order to establish relationships between calendar zones, our approach takes into account not only the calendar values involved in adverbial units but also the semantics of prepositional phrases involved in these units. Through a first experiment, we show how Information Retrieval systems could benefit from indexing calendar expressions in texts, defining relevance scores that combine keywords and temporal ranking models. © 2011 Springer-Verlag Berlin Heidelberg. Source

Amardeilh F.,Mondeca | Kraaij W.,TNO | Spitters M.,TNO | Versloot C.,TNO | Yurtsever S.,ATOS
Proceedings - 2013 European Intelligence and Security Informatics Conference, EISIC 2013 | Year: 2013

Domain ontologies are a central component in the Virtuoso demonstrator, a system that captures, analyzes and aggregates open news sources in order to achieve an information position that supports complex decision processes in the context of border control. However, maintenance of such an ontology is a challenging task. We demonstrate a text processing pipeline that supports domain experts in maintaining the domain ontology. The system facilitates the maintenance by generating candidate concepts that should be added to the ontology. Some initial tests have been carried out with filtering candidate concepts from a domain specific news feed. © 2013 IEEE. Source

Faucher C.,University of La Rochelle | Teissedre C.,French National Center for Scientific Research | Teissedre C.,Mondeca | Lafaye J.-Y.,University of La Rochelle | Bertrand F.,University of La Rochelle
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2010

The objectives of this paper are to present, describe, and explain the foundations and the functionalities of a temporal knowledge acquisition and modeling solution workflow, which aims at acquiring temporal knowledge from texts in order to populate a constrained object model. We are using several models for temporal data, one of which is generic and employed as a pivot model between a linguistic representation and a calendar representation. The approach we propose is generic and has been tested against a real use case, in which input data is made of temporal properties defining when a given location (a theater, a restaurant, a shopping center, etc.) is open or closed. Most expressions entered are expressed in intension. Our models provide a core support to the system that linguistically analyses data entries, transforms them into extensive calendar information and allow users to control the quality of the system's interpretation. © 2010 Springer-Verlag. Source

Garcia-Santa N.,Expert System | Atemezing G.A.,Mondeca | Villazon-Terrazas B.,Expert System
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015

Developing ontologies, by reusing already available and wellknown ontologies, is commonly acknowledge to play a crucial role to facilitate inclusion and expansion of the Web of Data. Some recommendations exist to guide ontologists in ontology engineering, but they do not provide guidelines on how to reuse vocabularies at low fine grained, i.e., reusing specific classes and properties. Moreover, it is still hard to find a tool that provides users with an environment to reuse terms. This paper presents ProtégéLOV, a plugin for the ontology editor Protégé, that combines the access to the Linked Open Vocabularies (LOV) during ontology modeling. It allows users to search a term in LOV and provides three actions if the term exists: (i) replace the selected term in the current ontology; (ii) add the rdfs:subClassOf or rdfs:subPropertyOf axiom between the selected term and the local term; and (iii) add the owl:equivalentClass or owl:equivalentProperty between the selected term and local term. Results from a preliminary user study indicate that ProtégéLOV does provide an intuitive access and reuse of terms in external vocabularies. © Springer International Publishing Switzerland 2015. Source

Baker T.,The Dublin Core Metadata Initiative - DCMI | Vandenbussche P.-Y.,Fujitsu Limited | Vatant B.,Mondeca
Library Hi Tech | Year: 2013

Purpose: The paper seeks to analyze the health of the vocabulary ecosystem in terms of requirements, addressing its various stakeholders such as maintainers of linked open vocabularies, linked data providers who use those vocabularies in their data and memory institutions which, it is hoped, will eventually provide for the long-term preservation of vocabularies. Design/methodology/approach: This paper builds on requirements formulated more tersely in the DCMI generic namespace policy for RDF vocabularies. The examination of requirements for linked open vocabularies focuses primarily on property-and-class vocabularies in RDFS or OWL (sometimes called metadata element sets), with some consideration of SKOS concept schemes and Dublin Core application profiles. It also discusses lessons learned through two years of development of the linked open vocabularies (LOV), of which main features and key findings are described. Findings: Key findings about the current practices of vocabulary managers regarding metadata, policy and versioning are presented, as well as how such practices can be improved to ensure better discoverability and usability of RDF vocabularies. The paper presents new ways to assess the links and dependencies between vocabularies. It also stresses the necessity and importance of a global governance of the ecosystem in which vocabulary managers, standard bodies, and memory institutions should engage. Research limitations/implications: The current paper is focused on requirements related to a single type of vocabulary but could and should be extended to other types such as thesauri, classifications, and other semantic assets. Practical implications: Practical technical guidelines and social good practices are proposed for promotion in the Vocabulary Ecosystem (for example, by vocabulary managers). Originality/value: This paper brings together the research and action of several important actors in the vocabulary management and governance field, and is intended to be the basis of a roadmap for action presented at the Dublin Core conference of September 2013 in Lisbon (DC 2013). © Emerald Group Publishing Limited. Source

Discover hidden collaborations