Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties

In this paper, we present an in-depth investigation of the linguistic knowledge encoded by the transformer models currently available for the Italian language. In particular, we investigate how the complexity of two different architectures of probing models affects the performance of the Transformer...

Full description

Bibliographic Details
Main Authors: Alessio Miaschi, Gabriele Sarti, Dominique Brunato, Felice Dell’Orletta, Giulia Venturi
Format: Article
Language:English
Published: Accademia University Press 2022-07-01
Series:IJCoL
Online Access:http://journals.openedition.org/ijcol/965
_version_ 1811214243970678784
author Alessio Miaschi
Gabriele Sarti
Dominique Brunato
Felice Dell’Orletta
Giulia Venturi
author_facet Alessio Miaschi
Gabriele Sarti
Dominique Brunato
Felice Dell’Orletta
Giulia Venturi
author_sort Alessio Miaschi
collection DOAJ
description In this paper, we present an in-depth investigation of the linguistic knowledge encoded by the transformer models currently available for the Italian language. In particular, we investigate how the complexity of two different architectures of probing models affects the performance of the Transformers in encoding a wide spectrum of linguistic features. Moreover, we explore how this implicit knowledge varies according to different textual genres and language varieties.
first_indexed 2024-04-12T06:01:13Z
format Article
id doaj.art-0d3569da32f04598b0dad3c104586778
institution Directory Open Access Journal
issn 2499-4553
language English
last_indexed 2024-04-12T06:01:13Z
publishDate 2022-07-01
publisher Accademia University Press
record_format Article
series IJCoL
spelling doaj.art-0d3569da32f04598b0dad3c1045867782022-12-22T03:45:02ZengAccademia University PressIJCoL2499-45532022-07-018110.4000/ijcol.965Probing Linguistic Knowledge in Italian Neural Language Models across Language VarietiesAlessio MiaschiGabriele SartiDominique BrunatoFelice Dell’OrlettaGiulia VenturiIn this paper, we present an in-depth investigation of the linguistic knowledge encoded by the transformer models currently available for the Italian language. In particular, we investigate how the complexity of two different architectures of probing models affects the performance of the Transformers in encoding a wide spectrum of linguistic features. Moreover, we explore how this implicit knowledge varies according to different textual genres and language varieties.http://journals.openedition.org/ijcol/965
spellingShingle Alessio Miaschi
Gabriele Sarti
Dominique Brunato
Felice Dell’Orletta
Giulia Venturi
Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties
IJCoL
title Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties
title_full Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties
title_fullStr Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties
title_full_unstemmed Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties
title_short Probing Linguistic Knowledge in Italian Neural Language Models across Language Varieties
title_sort probing linguistic knowledge in italian neural language models across language varieties
url http://journals.openedition.org/ijcol/965
work_keys_str_mv AT alessiomiaschi probinglinguisticknowledgeinitalianneurallanguagemodelsacrosslanguagevarieties
AT gabrielesarti probinglinguisticknowledgeinitalianneurallanguagemodelsacrosslanguagevarieties
AT dominiquebrunato probinglinguisticknowledgeinitalianneurallanguagemodelsacrosslanguagevarieties
AT felicedellorletta probinglinguisticknowledgeinitalianneurallanguagemodelsacrosslanguagevarieties
AT giuliaventuri probinglinguisticknowledgeinitalianneurallanguagemodelsacrosslanguagevarieties