Event · Etruscan: the readable language that cannot be fully translated

Vico and Spanakis release the Larth dataset and the first Etruscan machine-translation benchmark

In September 2023, at the Ancient Language Processing workshop, the first open Etruscan corpus for natural language processing appears: 7,139 inscriptions, 2,891 with English translations, and translation models scoring BLEU 10.1.

Happened to 1

Etruscan: the readable language that cannot be fully translated

MYS

Cites 1

Larth: Dataset and Machine Translation for Etruscan

SRC

Cite it

Mysteries Department. (2026). Vico and Spanakis release the Larth dataset and the first Etruscan machine-translation benchmark [Event EVT0DEDVF7RTYQW]. https://mysteriesdept.com/r/EVT0DEDVF7RTYQW

@misc{mysteriesdept_evt0dedvf7rtyqw,
  title     = {Vico and Spanakis release the Larth dataset and the first Etruscan machine-translation benchmark},
  author    = {{Mysteries Department}},
  year      = {2026},
  note      = {Event, catalog key EVT0DEDVF7RTYQW},
  url       = {https://mysteriesdept.com/r/EVT0DEDVF7RTYQW}
}