FAIRTAG4ALL

About the Tutorial

FAIRTAG4ALL is dedicated to the design, implementation, and use of FAIR-compliant terminological databases in the era of Generative AI. Building on the FAIR terminology approach and the FAIRterm 2.0 platform, the tutorial introduces methods for integrating structured terminological resources into Terminology-Augmented Generation (TAG) workflows for Large Language Models.

Participants will learn how to design terminology resources compliant with ISO and W3C standards, implement TBX-to-RDF workflows, and connect terminological data to Linked Data infrastructures ensuring interoperability, transparency, and explainability of generative systems.

The tutorial blends theoretical discussion with hands-on exercises using FAIRterm 2.0, bridging established terminology practice and responsible AI development.

Target Audience

The tutorial is intended for an interdisciplinary audience including: terminologists, computational linguists, NLP researchers, digital humanists, translators, data stewards, and developers of language resources.

Basic familiarity with linguistic data modeling is beneficial. Knowledge of XML, RDF, or ISO terminology standards is helpful but not required. For the TAG component, a general understanding of Large Language Models is assumed; no programming experience is necessary since demonstrations are tool-based.

Tutorial Program

Module 1 – Conceptual Foundations: FAIR Terminology and Open Science (45 min)

• FAIR principles for terminological data.
• Open infrastructures supporting linguistic resources (CLARIN, ELG).
• Reproducible and shareable terminology practices.

Module 2 – Standardization and Data Modeling (45 min)

• ISO TC/37 standards for terminology management.
• Terminological data models: TBX, RDF, Linked Data.
• TBX-to-RDF workflows and semantic interoperability.
• FAIRterm 2.0 data model and exports.

Coffee Break (30 min)

Module 3 – Practical Session: Building FAIR Databases (45 min)

• Creating multilingual terminological datasets with FAIRterm 2.0.
• Metadata design and licensing under FAIR conditions.
• Exporting data (TBX, XML, RDF) and linking external resources.

Module 4 – Terminology-Augmented Generation (45 min)

• Foundations of TAG.
• Using terminological resources to enhance LLM reliability.
• Terminology-driven prompting and RAG workflows.
• Explainability and responsible AI practices.

Presenters

Giorgio Maria Di Nunzio

Department of Information Engineering, University of Padua, Italy

Federica Vezzani

Department of Linguistic and Literary Studies, University of Padua, Italy

Tutorial Materials

Slides, datasets, exercises, and access instructions for the FAIRterm 2.0 platform will be released openly prior to the tutorial. All resources will be distributed under open licenses to facilitate reuse and reproducibility.

Reading List

The following references provide conceptual background and methodological foundations for the topics addressed in the FAIRTAG4ALL tutorial.

FAIR Principles and Interoperability

These works provide the conceptual and technical background for designing FAIR-compliant terminological resources and interoperable infrastructures.

• Wilkinson, M. D. et al. (2016). The FAIR Guiding Principles for Scientific Data Management and Stewardship.
  Link to paper
• Vezzani, F. (2022). Terminologie numérique : conception, représentation et gestion.
  Link to book
• Vezzani, F. & Di Nunzio, G. M. (2020). Methodology for the Standardization of Terminological Resources: Design of the TriMED Database to support multi-register medical communication.
  Link to paper
• Vezzani, F. & Di Nunzio, G. M. (2020). On the Formal Standardization of Terminology Resources: The TriMED Case Study.
  Link to paper
• Vezzani, F., Di Nunzio, G. M., Costa, R. (2023). ISO standards for terminology resources management: Are they FAIR enough?
  Link to paper

Terminology-Augmented Generation (TAG) and Generative AI

These papers explore the integration of controlled terminological knowledge into Large Language Models and generative AI workflows.

• Di Nunzio, G. M. (2025). Terminology-Augmented Generation (TAG): Foundations, Use Cases, and Evaluation Paths.
  Link to paper
• Lackner, A., Vega-Wilson, A., & Lang, C. (2025). An Evaluation of Terminology-Augmented Generation (TAG) and Various Terminology Formats for the Translation Use Case.
  Link to paper

Terminology Theory, Linked Data, and Interoperability Models

These readings focus on theoretical and representational frameworks for the integration of terminological, lexical, and semantic data in digital environments.

• Cimiano, P., Chiarcos, C., McCrae, J. P., & Gracia, J. (2020). Linguistic Linked Data. Representation, Generation and Applications
  Link to book
• Bellandi, A., Di Nunzio, G. M., Piccini, S., Vezzani, F. (2024). LemonizeTBX: Design and Implementation of a New Converter from TBX to OntoLex-Lemon
  Link to paper

Venue

FAIRTAG4ALL will be delivered as part of the official tutorial program of LREC 2026 in Palma de Mallorca, Spain. Information on registration, accommodation, and travel can be found on the main conference website: https://lrec2026.info/.

Contact

For information about the tutorial please contact:
giorgiomaria.dinunzio@unipd.it