2017 Fiscal Year Final Research Report
Semantic similarity using human associative knowledge
| Project/Area Number |
26540144
|
|---|
| Research Category |
Grant-in-Aid for Challenging Exploratory Research
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Kansei informatics
|
|---|
| Research Institution | Waseda University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2014-04-01 – 2018-03-31
|
| Keywords | 想起関係 / 意味的類似度 / 意味関係 / 分散表現 / マルチモーダル / ソーシャルタグ |
|---|
| Outline of Final Research Achievements |
Developing an appropriate computational mechanism of semantic similarity between linguistic expressions is an important subject for both engineering applications and cognitive science. In this research project, by focusing on evocation relationships of semantic concepts that human beings implicitly organize in their brains, new computational methods for measuring semantic similarity between lexical concepts and for classifying potential semantic relationships between them have been studied. These methods utilize machine learning techniques, including deep neural networks, for integrating linguistic features with image-originated perceptual features, as well as social implications/meanings derived from social image tags. Our methods achieved nealy state-of-the-art results in semantic similarity/relatedness tasks and classification of lexical semantic relations. These results have been discussed in several international and domestic conferences.
|
| Free Research Field |
自然言語処理，意味コンピューティング
|
