Enregistré dans:
Détails bibliographiques
Auteurs principaux: Luisto, Rami, Petäinen, Liisa, Grönholm, Tommi, Böhm, Jan, Ahtiainen, Maarit, Lilja, Tomi, Pölönen, Ilkka, Äyrämö, Sami
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2604.14815
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910135069179904
author Luisto, Rami
Petäinen, Liisa
Grönholm, Tommi
Böhm, Jan
Ahtiainen, Maarit
Lilja, Tomi
Pölönen, Ilkka
Äyrämö, Sami
author_facet Luisto, Rami
Petäinen, Liisa
Grönholm, Tommi
Böhm, Jan
Ahtiainen, Maarit
Lilja, Tomi
Pölönen, Ilkka
Äyrämö, Sami
contents In NLP classification tasks where little labeled data exists, domain fine-tuning of transformer models on unlabeled data is an established approach. In this paper we have two aims. (1) We describe our observations from fine-tuning the Finnish BERT model on Finnish medical text data. (2) We report on our attempts to predict the benefit of domain-specific pre-training of Finnish BERT from observing the geometry of embedding changes due to domain fine-tuning. Our driving motivation is the common\situation in healthcare AI where we might experience long delays in acquiring datasets, especially with respect to labels.
format Preprint
id arxiv_https___arxiv_org_abs_2604_14815
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Domain Fine-Tuning FinBERT on Finnish Histopathological Reports: Train-Time Signals and Downstream Correlations
Luisto, Rami
Petäinen, Liisa
Grönholm, Tommi
Böhm, Jan
Ahtiainen, Maarit
Lilja, Tomi
Pölönen, Ilkka
Äyrämö, Sami
Computation and Language
In NLP classification tasks where little labeled data exists, domain fine-tuning of transformer models on unlabeled data is an established approach. In this paper we have two aims. (1) We describe our observations from fine-tuning the Finnish BERT model on Finnish medical text data. (2) We report on our attempts to predict the benefit of domain-specific pre-training of Finnish BERT from observing the geometry of embedding changes due to domain fine-tuning. Our driving motivation is the common\situation in healthcare AI where we might experience long delays in acquiring datasets, especially with respect to labels.
title Domain Fine-Tuning FinBERT on Finnish Histopathological Reports: Train-Time Signals and Downstream Correlations
topic Computation and Language
url https://arxiv.org/abs/2604.14815