SciBERT: Pretrained Contextualized Embeddings for Scientific Text

Explore this paper's citation graph

Summary

SciBERT leverages unsupervised pretraining on a large multi-domain corpus of scientific publications to improve performance on downstream scientific NLP tasks and demonstrates statistically significant improvements over BERT.

Type
article
Published
2019-03-26
Cited by
227
References
29

Keywords

Computer science, Suite, Dependency grammar, Natural language processing, Variety (cybernetics)

References

Cited by

Related papers