Skip to Main content Skip to Navigation
Poster communications

Re-train or Train from Scratch? Comparing Pre-training Strategies of BERT in the Medical Domain

Abstract : BERT models used in specialized domains all seem to be the result of a simple strategy: initializing with the original BERT then resuming pre-training on a specialized corpus. This method yields rather good performance (e.g. BioBERT (Lee et al., 2020), SciBERT (Beltagy et al., 2019), BlueBERT (Peng et al., 2019)). However, it seems reasonable to think that training directly on a specialized corpus, using a specialized vocabulary, could result in more tailored embeddings and thus help performance. To test this hypothesis, we train BERT models from scratch using many configurations involving general and medical corpora. Based on evaluations using four different tasks, we find that the initial corpus only has a weak influence on the performance of BERT models when these are further pre-trained on a medical corpus.
Document type :
Poster communications
Complete list of metadata

https://hal-cea.archives-ouvertes.fr/cea-03745341
Contributor : Contributeur MAP CEA Connect in order to contact the contributor
Submitted on : Thursday, August 4, 2022 - 9:22:20 AM
Last modification on : Saturday, August 6, 2022 - 3:26:35 AM

File

2022.lrec-1.281.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : cea-03745341, version 1

Citation

Hicham El Boukkouri, Olivier Ferret, Thomas Lavergne, Pierre Zweigenbaum. Re-train or Train from Scratch? Comparing Pre-training Strategies of BERT in the Medical Domain. LREC 2022, Jun 2022, Marseille, France. pp.2626-2633, 2022, LREC 2022. ⟨cea-03745341⟩

Share

Metrics

Record views

0

Files downloads

0