We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Text Mining Drug/Chemical-Protein Interactions using an Ensemble of BERT and T5 Based Models

Abstract: In Track-1 of the BioCreative VII Challenge participants are asked to identify interactions between drugs/chemicals and proteins. In-context named entity annotations for each drug/chemical and protein are provided and one of fourteen different interactions must be automatically predicted. For this relation extraction task, we attempt both a BERT-based sentence classification approach, and a more novel text-to-text approach using a T5 model. We find that larger BERT-based models perform better in general, with our BioMegatron-based model achieving the highest scores across all metrics, achieving 0.74 F1 score. Though our novel T5 text-to-text method did not perform as well as most of our BERT-based models, it outperformed those trained on similar data, showing promising results, achieving 0.65 F1 score. We believe a text-to-text approach to relation extraction has some competitive advantages and there is a lot of room for research advancement.
Comments: Submission to the BioCreative VII challenge, Track-1
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2111.15617 [cs.CL]
  (or arXiv:2111.15617v1 [cs.CL] for this version)

Submission history

From: Hoo Chang Shin [view email]
[v1] Tue, 30 Nov 2021 18:14:06 GMT (4297kb,D)

Link back to: arXiv, form interface, contact.