2020 Fiscal Year Annual Research Report
Novel Technologies for Improving Comprehension and Utilization of Historical Knowledge
| Project/Area Number |
17H01828
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
Adam Jatowt 京都大学, 情報学研究科, 特定准教授 (00415861)
|
|---|
| Project Period (FY) |
2017-04-01 – 2021-03-31
|
| Keywords | digital archives / computational history / temporal information / question answering |
|---|
| Outline of Annual Research Achievements |
We developed effective method for answering questions in long-term news archives. For any arbitrary question about the past, our method provides often a correct answer. This is a novel kind of question answering where temporal information needs to be carefully considered. We have constructed a dataset of multiple answers and their questions for this research which is open. We have also published a full paper at ECIR2020 (core A) and journal paper at IRJ (IF 2.2). The ECIR paper was nominated for industry impact award paper at ECIR2020.
We have also completed the research on retrieval of interesting information from the long-term archives that appeared in ECIR2021 conference as a full paper.
Lastly, the novel research on multiple timeline summarization has been also completed with proposed model having good results.
|
| Research Progress Status |
令和2年度が最終年度であるため、記入しない。
|
| Strategy for Future Research Activity |
令和2年度が最終年度であるため、記入しない。
|
