Does constituency analysis enhance domain-specific pre-trained BERT models for relation extraction?

Anfu Tang (1, 2) , Louise Deléger (1) , Robert Bossy (1) , Pierre Zweigenbaum (3) , Claire Nédellec (1)

1 MaIAGE - Mathématiques et Informatique Appliquées du Génome à l'Environnement [Jouy-En-Josas]
2 LISN - Laboratoire Interdisciplinaire des Sciences du Numérique
3 ILES - Information, Langue Ecrite et Signée

Recently many studies have been conducted on the topic of relation extraction. The DrugProt track at BioCreative VII provides a manually-annotated corpus for the purpose of the development and evaluation of relation extraction systems, in which interactions between chemicals and genes are studied. We describe the ensemble system that we used for our submission, which combines predictions of fine-tuned bioBERT, sciBERT and const-bioBERT models by majority voting. We specifically tested the contribution of syntactic information to relation extraction with BERT. We observed that adding constituentbased syntactic information to BERT improved precision, but decreased recall, since relations rarely seen in the train set were less likely to be predicted by BERT models in which the syntactic information is infused. Our code is available online [https://github.com/Maple177/drugprot-relation-extraction].

Mots clés: Relation Extraction; Deep Learning; Transformers; BERT (Bidirectional Encoder Representations from Transformers); constituency analysis; chemical-gene interactions

Domaines: Bio-informatique [q-bio.QM]; Apprentissage [cs.LG]; Informatique et langage [cs.CL]; Intelligence artificielle [cs.AI]

Anfu Tang, Louise Deléger, Robert Bossy, Pierre Zweigenbaum, Claire Nédellec. Does constituency analysis enhance domain-specific pre-trained BERT models for relation extraction?. BioCreative VII Challenge Evaluation Workshop, Nov 2021, on-line, Spain. ⟨hal-03447774⟩