Wide-coverage event extraction from biomedical texts
Project/Area Number |
25730129
|
Research Category |
Grant-in-Aid for Young Scientists (B)
|
Allocation Type | Multi-year Fund |
Research Field |
Intelligent informatics
|
Research Institution | Toyota Technological Institute (2014-2015) The University of Tokyo (2013) |
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) |
2013-04-01 – 2016-03-31
|
Project Status |
Completed (Fiscal Year 2015)
|
Budget Amount *help |
¥3,640,000 (Direct Cost: ¥2,800,000、Indirect Cost: ¥840,000)
Fiscal Year 2015: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2014: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2013: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
|
Keywords | 事象抽出 / 複数コーパス / 半教師あり学習 / 教師なし学習 / 生命医学文献 / 高被覆 |
Outline of Final Research Achievements |
Biomedical event extraction systems often employ supervised machine learning approaches to learn from an annotated corpus. Such systems, however, can only extract limited types of events due to the limited annotated information, and it is costly to annotate a large amount of texts to cover a wide variety of event types. To deal with these problems, we propose and evaluate a method to build a wide-coverage model from several corpora and a method to find annotation candidates from unannotated texts. We show that the learning from several corpora can improve the event extraction performance, and we also present the possibility to use unannotated texts in event extraction.
|
Report
(4 results)
Research Products
(7 results)