Experiments on transfer learning architectures for biomedical relation extraction
This work provides an incremental improvement in biomedical relation extraction for researchers and practitioners working with scientific texts.
This paper explores BERT-based architectures and transfer learning strategies for biomedical relation extraction, achieving state-of-the-art performance on two corpora with absolute improvements of 1.73% on ChemProt and 32.77% on PGxCorpus using a BERT-segMCNN with finetuning.
Relation extraction (RE) consists in identifying and structuring automatically relations of interest from texts. Recently, BERT improved the top performances for several NLP tasks, including RE. However, the best way to use BERT, within a machine learning architecture, and within a transfer learning strategy is still an open question since it is highly dependent on each specific task and domain. Here, we explore various BERT-based architectures and transfer learning strategies (i.e., frozen or fine-tuned) for the task of biomedical RE on two corpora. Among tested architectures and strategies, our *BERT-segMCNN with finetuning reaches performances higher than the state-of-the-art on the two corpora (1.73 % and 32.77 % absolute improvement on ChemProt and PGxCorpus corpora respectively). More generally, our experiments illustrate the expected interest of fine-tuning with BERT, but also the unexplored advantage of using structural information (with sentence segmentation), in addition to the context classically leveraged by BERT.