IMPOLS at EVALITA 2026: Overview of the IMPOLS Task

Lorenzo Gregori, Walter Paci, Valentina Saccone

Keywords: political speech, implicit content, EVALITA 2026

TL;DR: IMPOLS report at EVALITA 2026

Abstract: This paper describes the IMPOLS task presented at EVALITA 2026 competition. The task regards the automatic detection and classification of implicit contents in political speeches, specifically focusing on the uses of implicit expressions that are potentially manipulative. The task is divided in three subtasks: (1) implicit content detection, (2) implicit classification (implicature and presupposition), and (3) implicature type classification. IMPOLS is monolingual (Italian) multimodal task: provided data included audio and transcription of speech excerpts, along with some discourse metadata. Two teams participated at IMPOLS, obtaining a macro F1 score of 0.95 on subtask 1, 0.88 on subtask 2, and 0.64 on subtask 3.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 1