This paper presents an overview of the International Workshop on Definition and Term Extraction Challenge (DETECH) 2026, held in conjunction with Multilingual Digital Terminology Today (MDTT) 2026. DETECH 2026 focuses on automatic term extraction and definition generation in the biomedical domain, with particular attention to the gut-brain interplay and to two related subdomains: mental health and Parkinson’s disease. The challenge is organized around two tasks: Task A, devoted to the extraction of single-word and multi-word biomedical terms from English PubMed abstracts, and Task B, devoted to concept assignment and natural language definition generation. This overview introduces the scientific motivation of the challenge, describes its task setting and evaluation framework, and summarizes the accepted contributions. The papers presented at DETECH 2026 cover a broad methodological spectrum, including symbolic term extraction, BERT-based sequence labelling, biomedical transformer models, customized generative AI, retrieval-augmented generation, ontology-enhanced prompting, and theory-driven approaches to terminological meaning. Taken together, the contributions show that term extraction and definition generation should be treated as connected knowledge-oriented tasks requiring terminological adequacy, biomedical grounding, explainability, and expert-informed evaluation.
Overview of the International Workshop on Definition and Term Extraction Challenge (DETECH) 2026
Vezzani F.;Di Nunzio G. M.
;Bonato V.;Silvello G.
2026
Abstract
This paper presents an overview of the International Workshop on Definition and Term Extraction Challenge (DETECH) 2026, held in conjunction with Multilingual Digital Terminology Today (MDTT) 2026. DETECH 2026 focuses on automatic term extraction and definition generation in the biomedical domain, with particular attention to the gut-brain interplay and to two related subdomains: mental health and Parkinson’s disease. The challenge is organized around two tasks: Task A, devoted to the extraction of single-word and multi-word biomedical terms from English PubMed abstracts, and Task B, devoted to concept assignment and natural language definition generation. This overview introduces the scientific motivation of the challenge, describes its task setting and evaluation framework, and summarizes the accepted contributions. The papers presented at DETECH 2026 cover a broad methodological spectrum, including symbolic term extraction, BERT-based sequence labelling, biomedical transformer models, customized generative AI, retrieval-augmented generation, ontology-enhanced prompting, and theory-driven approaches to terminological meaning. Taken together, the contributions show that term extraction and definition generation should be treated as connected knowledge-oriented tasks requiring terminological adequacy, biomedical grounding, explainability, and expert-informed evaluation.
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
