CNRS dépasser les frontières

facilite l’accès aux résultats issus des différents champs de la recherche mondiale

Terminology and Natural Language Processing

Partager

The Inist’s engineers provide skills to best respond to the requirements of researchers, laboratories, research groups or institutions :

  • normalized, validated and interoperable terminological resources.

These resources are in the form of thesauruses, ontoterminologies or ontologies and respond to requirements linked to visibility, indexing, annotation, information searches, knowledge structuring, analysis of textual contents, technical copyrighting, support for decision-making, etc.
Some of these achievements can be consulted on the site : http://terminologies.inist.fr/
They can also be downloaded from the "Terminologies" page of the Ortolang platform :
https://www.ortolang.fr/market/terminologies

  • analysis of textual corpora to improve the processes of information searches, extraction or data handling. These processes are particularly used for the purposes of scientific steering or research which require data management involving published knowledge (text and data mining), terminological data monitoring, etc.

The services of the Terminology-NLP team are provided in collaboration with users and are based on current norms and standards along with terminology management tools and natural language processing.

Two structuring projects in response to external and internal requirements are underway :

  • LOTERRE will offer a consultation service (human visualization) and online display (using Linked Data methods) of terminological resources which will also integrate the team’s service offer ;
  • LegoTAL aims to create an open, perennial text-mining infrastructure to enable Inist engineers to develop, execute, manage and share Natural Language Processing (NLP) chains which are applied to document collections.

Inist decided to stop developments linked to the LegoTAL project as of September 1st 2017, opting instead to use the OpenMinTeD platform. The Inist is therefore now involved in the Visa TM project (Towards an advanced text-mining services infrastructure) supported by the French Ministry of Higher Education and Research which aims to demonstrate the potential of results from the OpenMinTeD project for usage which would benefit French scientific communities.

The following partnership projects demonstrate the recognition of the Inist’s expertise in the production, processing, dissemination and sharing of approved, enriched terminological resources which are part of the Linked Data sphere :

  • BnF : enriching Rameau with specialist terminologies ;
  • L’ETP au fil des mots [1] : partnership with the Pôle Lorrain d’Education Thérapeutique du Patient (Lorraine Therapeutic Education for Patients Centre) to develop a terminology for the Therapeutic Education for Patients ;
  • FRANTIQ : alignment of the PACTOLS micro-thesaurus "Lieux" with GeoNames ;
  • SémanDiv GDR (The Semantics of Biodiversity Research Network) : creation of a thesaurus and ontology in the field of biodiversity.