Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

BioBERT: a pre-trained biomedical language representation model for biomedical text mining #29

Open
hideyoshikato opened this issue Feb 16, 2020 · 0 comments

Comments

@hideyoshikato
Copy link
Owner

0. 論文

https://arxiv.org/abs/1901.08746

1. 要約

問題:NLPを生物医学文書に直接適用すると,一般的なドメインコーパスから生物医学コーパスへの単語分布のシフトによって,満足のいく結果が得られない
対処法:pretrainされたBERTによってドメイン適応させる
結果:biomedical named entity recognition1,biomedical relation extraction,biomedical question answering3

2. 先行研究との差異

BERTを生物医学コーパスのドメイン適応させた点

3. 技術や手法のポイント

生物医学コーパスでpre-training

4. 有効性の検証方法

使用データ:生物医学コーパス
対戦相手:BERT (Devlin et al., 2019),BioBERT (þPubMed) ,BioBERT (þPMC),BioBERT (þPubMed þ PMC)
評価指標:F1スコア,MRR

5. 議論

6. 次に読む論文

7. メモ

*1 biomedical named entity recognition・・・文章の中から生物医学専門用語を認識して文章を分類するタスク
https://medium.com/@Hironsan/biomedical-named-entity-recognition-5404c4d15c61
*2 biomedical relation extraction・・・生物医学関係を可視化するタスク
https://www.sciencedirect.com/science/article/pii/S1532046418300534
*3 biomedical question answering・・・医学質問への応答を作成するタスク.
https://www.aclweb.org/anthology/W18-5308/

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant