| Project/Area Number |
17K12741
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Toyota Technological Institute |
|---|
Principal Investigator |
Miwa Makoto 豊田工業大学, 工学(系)研究科(研究院), 准教授 (00529646)
|
|---|
| Project Period (FY) |
2017-04-01 – 2020-03-31
|
| Project Status |
Completed (Fiscal Year 2019)
|
| Budget Amount *help |
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2019: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2018: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2017: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
|
|---|
| Keywords | 薬物間相互作用 / DrugBank / 関係抽出 / 深層学習 / 表現学習 / BERT / 畳み込みニューラルネットワーク / グラフニューラルネットワーク / グラフ / データベース / 言語処理 |
|---|
| Outline of Final Research Achievements |
To find important information from increasing the medical and pharmaceutical literature, there is a need for a method that automatically extracts relationship information of drugs, which is essential to R&D and use of drugs． In this research, we improved the conventional extraction method, which used to focus only on textual information, by effectively using the database information that is already discovered and organized. We have developed deep learning methods that utilize information on chemical structures in the database and a large amount of unannotated texts, and improved the conventional method, which showed around 70% in extraction performance, by more than 10% points.
|
| Academic Significance and Societal Importance of the Research Achievements |
従来の日本語や英語などの自然言語を対象とした情報抽出の研究では，言語情報のみを利用したものがほとんどであり，データベース情報などは「データベースにあるかどうか」などの特徴として補助的に使われる程度であった．本研究成果は，データベース上に含まれる化学式などの単純には言語と結びつかないような情報を，深層学習を用いて自然言語からの情報抽出に利用可能にし，さらにその情報を使うことで従来手法の精度を向上できたことに意義がある．
|
