| Project/Area Number |
18K19841
|
|---|
| Research Category |
Grant-in-Aid for Challenging Research (Exploratory)
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Medium-sized Section 62:Applied informatics and related fields
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
Jatowt Adam 京都大学, 情報学研究科, 特定准教授 (00415861)
|
|---|
| Project Period (FY) |
2018-06-29 – 2021-03-31
|
| Project Status |
Completed (Fiscal Year 2020)
|
| Budget Amount *help |
¥6,240,000 (Direct Cost: ¥4,800,000、Indirect Cost: ¥1,440,000)
Fiscal Year 2019: ¥2,210,000 (Direct Cost: ¥1,700,000、Indirect Cost: ¥510,000)
Fiscal Year 2018: ¥4,030,000 (Direct Cost: ¥3,100,000、Indirect Cost: ¥930,000)
|
|---|
| Keywords | news archives / language change / question answering / document archives / document archive / text analysis |
|---|
| Outline of Final Research Achievements |
We have developed approaches for answering question in long-term news archives. Our method can answer arbitrary user query about the past by extracting content from news articles that were published long time ago. This is challenging task due to many repeating and periodical events. For this research we have built a small dataset of 1k questions that contain answers. Our approaches are unsupervised and are based on estimating question time scope and then on retrieving content from news archives that fall within or relates to that time scope. After search result reranking using special module answers are produced from individual pages and are aggregated. During the research progress we found out several important observations such as how to find the time scope in the best way or how to combine document relevance with temporal relevance of documents. We have published a paper in core A ranked conference and a journal paper that was invited from that conference submission.
|
| Academic Significance and Societal Importance of the Research Achievements |
Based on the proposed approaches users can send questions to the past and obtain detailed information without the need to manually search and browse large news article archives. Journalists, historians and anyone who wishes to obtain answers about the past can benefit from this research.
|
