[apologies for cross-posting]
Call for Participation (CfP)
ATE-IT: Automatic Term Extraction — Italian Testbed
Shared Task @ EVALITA 2026
Overview
Automatic Term Extraction (ATE) is a core task in Natural Language Processing and Terminology, supporting applications such as knowledge acquisition, ontology building, and domain-specific lexicon development. ATE-IT is the first large-scale shared task dedicated to Italian. The task focuses on institutional texts from the waste management domain, providing a real-world testbed for evaluating methods from diverse methodological perspectives. ATE-IT aims not only to benchmark extraction performance, but also to advance research in ATE and terminology.
The shared task is divided into two subtasks:
Subtask A – Term Extraction: Identify relevant single-word and multi-word terms from domain texts.
Subtask B – Term Variant Clustering: Group extracted terms that refer to the same concept.
Participation
We welcome participation from students, researchers, and industry teams.
Up to two runs per subtask per team.
External resources are allowed but must be documented.
Manual runs are accepted but will not be ranked.
To register, please follow the instructions on the task website.
Evaluation
Subtask A: Micro-F1, Type-F1
Subtask B: BCubed-F1
Baselines and evaluation scripts will be provided
Submission of Technical Reports
Participants should write one technical report that will be published as official Proceedings of EVALITA 2026.
Details about the submission and publication will be provided soon.
In general, the report should contain:
The methodology and approach used
Experimental setup and results
Analysis and discussion of findings
Comparison with baseline methods
Timeline
Sept 22, 2025 — Training data release
Nov 21, 2025 — Registration deadline
Nov 24 – Dec 1, 2025 — Evaluation phase
Dec 15, 2025 — Results
Jan 9, 2026 — Participant report submission
Feb 7, 2026 — Reviews sent to participants
Feb 16, 2026 — Camera-ready report submission
Feb 26–27, 2026 — Workshop @ EVALITA (Bari)
Organizers
Nicola Cirillo, Università degli Studi di Salerno — nici...@unisa.it
Giorgio Maria Di Nunzio, Università degli Studi di Padova — giorgiomar...@unipd.it
Federica Vezzani, Università degli Studi di Padova — federica...@unipd.it
Contact
Email: ateit...@gmail.com
Discord: https://discord.gg/7MBGSJQzT7
ATE-IT website: https://nicolacirillo.github.io/ate-it/
EVALITA 2026 website: https://www.evalita.it/campaigns/evalita-2026/