Saved in:
Bibliographic Details
Main Authors: Buonocore, Tommaso Mario, Rancati, Simone, Parimbelli, Enea
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.06011
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909246542577664
author Buonocore, Tommaso Mario
Rancati, Simone
Parimbelli, Enea
author_facet Buonocore, Tommaso Mario
Rancati, Simone
Parimbelli, Enea
contents The development of domain-specific language models has significantly advanced natural language processing applications in various specialized fields, particularly in biomedicine. However, the focus has largely been on English-language models, leaving a gap for less-resourced languages such as Italian. This paper introduces Igea, the first decoder-only language model designed explicitly for biomedical text generation in Italian. Built on the Minerva model and continually pretrained on a diverse corpus of Italian medical texts, Igea is available in three model sizes: 350 million, 1 billion, and 3 billion parameters. The models aim to balance computational efficiency and performance, addressing the challenges of managing the peculiarities of medical terminology in Italian. We evaluate Igea using a mix of in-domain biomedical corpora and general-purpose benchmarks, highlighting its efficacy and retention of general knowledge even after the domain-specific training. This paper discusses the model's development and evaluation, providing a foundation for future advancements in Italian biomedical NLP.
format Preprint
id arxiv_https___arxiv_org_abs_2407_06011
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Igea: a Decoder-Only Language Model for Biomedical Text Generation in Italian
Buonocore, Tommaso Mario
Rancati, Simone
Parimbelli, Enea
Computation and Language
Artificial Intelligence
I.2.7; J.3
The development of domain-specific language models has significantly advanced natural language processing applications in various specialized fields, particularly in biomedicine. However, the focus has largely been on English-language models, leaving a gap for less-resourced languages such as Italian. This paper introduces Igea, the first decoder-only language model designed explicitly for biomedical text generation in Italian. Built on the Minerva model and continually pretrained on a diverse corpus of Italian medical texts, Igea is available in three model sizes: 350 million, 1 billion, and 3 billion parameters. The models aim to balance computational efficiency and performance, addressing the challenges of managing the peculiarities of medical terminology in Italian. We evaluate Igea using a mix of in-domain biomedical corpora and general-purpose benchmarks, highlighting its efficacy and retention of general knowledge even after the domain-specific training. This paper discusses the model's development and evaluation, providing a foundation for future advancements in Italian biomedical NLP.
title Igea: a Decoder-Only Language Model for Biomedical Text Generation in Italian
topic Computation and Language
Artificial Intelligence
I.2.7; J.3
url https://arxiv.org/abs/2407.06011