Controllable Natural Language Generation with Neural Networks
| Project/Area Number |
16H05873
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (A)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Nara Institute of Science and Technology |
|---|
Principal Investigator |
Neubig Graham 奈良先端科学技術大学院大学, 先端科学技術研究科, 客員准教授 (70633428)
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Project Status |
Completed (Fiscal Year 2018)
|
| Budget Amount *help |
¥22,620,000 (Direct Cost: ¥17,400,000、Indirect Cost: ¥5,220,000)
Fiscal Year 2018: ¥5,070,000 (Direct Cost: ¥3,900,000、Indirect Cost: ¥1,170,000)
Fiscal Year 2017: ¥8,840,000 (Direct Cost: ¥6,800,000、Indirect Cost: ¥2,040,000)
Fiscal Year 2016: ¥8,710,000 (Direct Cost: ¥6,700,000、Indirect Cost: ¥2,010,000)
|
|---|
| Keywords | 言語生成 / 機械翻訳 / ニューラルネットワーク / 機械学習 / 自然言語処理 / 人工知能 / 深層学習 / 自然言語生成 / 制御性 |
|---|
| Outline of Final Research Achievements |
Technology to generate sentences in natural language has various applications: machine translation, which translates between human languages, dialogue systems, which facilitate communication between humans and machines, and summarization, which converts longer chunks of content into short summaries. Recently, methods based on neural networks (NNs) have outperformed other more traditional models by a large margin, but they often also cause unpredictable and serious errors. In this project, we developed NN-based language generation technology where it is possible to control the results, maintaining the accuracy of NN-based methods while improving practical applicability.
|
| Academic Significance and Societal Importance of the Research Achievements |
それぞれの研究成果は、言語生成の性能向上、効率向上、制御性向上などを実現しており、機械翻訳・対話システム・自動要約などの幅広い分野で適応可能である。それぞれの応用において、誤りの削減や制御性の向上、効率の向上を実験により確認している。また、開発したソフトウェアを公開し、研究機関や企業などで広く利用されている。
|
Report
(4 results)
Research Products
(22 results)
