LLMs

Open-vocabulary Word Sense Disambiguation at Digital Latin III
Open-vocabulary Word Sense Disambiguation at Digital Latin III

I presented preliminary results on LLM-based Word Sense Disambiguation for Latin at the third Digital Latin workshop in Siena.

Jun 17, 2026

LLM-Driven, Open-Vocabulary Word Sense Disambiguation for Latin Digital Lexicography
LLM-Driven, Open-Vocabulary Word Sense Disambiguation for Latin Digital Lexicography

I presented Inspicio, an LLM-based pipeline that retrieves Open English WordNet senses for Latin words without a predefined list of candidates, at the third Digital Latin workshop in Siena.

Jun 16, 2026

LLMs and the humanities at the Festival dell'Economia di Trento
LLMs and the humanities at the Festival dell'Economia di Trento

Notes from my panel at the 2026 Festival dell'Economia di Trento, with Luca De Biase and Ersilia Vaudo.

May 25, 2026

Un futuro perfettamente prevedibile
Un futuro perfettamente prevedibile

A panel with Luca De Biase (Il Sole 24 Ore) and Ersilia Vaudo (European Space Agency) at the 21st Festival dell'Economia di Trento, organised by Il Sole 24 Ore. I talked about LLMs from the point of view of the humanities.

May 22, 2026

My visiting research stay at King's College London
My visiting research stay at King's College London

Notes on my H2IOSC-funded visiting research stay at the Department of Digital Humanities, King's College London.

Jan 22, 2026

Expanding PREMOVE: An LLM-Assisted Semantic Annotation of Preverbs Across Historical and Modern Languages
Expanding PREMOVE: An LLM-Assisted Semantic Annotation of Preverbs Across Historical and Modern Languages

A seminar on expanding PREMOVE, a diachronic dataset of Latin and Ancient Greek preverbs, with perception verbs and Italian motion verbs annotated through an LLM-assisted pipeline. Given during my H2IOSC-funded visiting research stay at King's College London.

Jan 20, 2026

Veras Audire et Reddere Voces — Prosodically‑Accurate Latin Text-to-Speech Corpus and Methodology
Veras Audire et Reddere Voces — Prosodically‑Accurate Latin Text-to-Speech Corpus and Methodology

An openly licensed, expertly validated corpus of Latin poetic audio that makes classical metre audible with modern large‑language‑model TTS. Drawing on Pedecerto’s machine‑readable scansions, we steered a general‑purpose model to recite 216 lines—Vergil Aeneid 1.1–100 (hexameter) and Ovid Heroides 1.1–116 (elegiac couplets)—with metrically faithful stress, yielding nearly 24 minutes of line‑aligned, classroom‑ready audio. This project was presented at the ACL 2025 Student Research Workshop (Vienna) and at CLiC‑it 2025 (Cagliari).

Jul 26, 2025