Development of methods for extractive and abstractive text summarization based on machine learning with large-scale data

2016 Fiscal Year Final Research Report

• Project/Area Number: 26280080
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: Intelligent informatics
• Research Institution: Tokyo Institute of Technology
• Principal Investigator: Takamura Hiroya 東京工業大学, 科学技術創成研究院, 准教授 (80361773)
• Co-Investigator(Kenkyū-buntansha): 笹野 遼平 東京工業大学, 精密工学研究所, 助教 (70603918)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 文書要約 / 機械学習 / 大規模データ / ニューラルネットワーク

Outline of Final Research Achievements

We developed a method for automatically creating a large-scale training dataset for text summarization as well as a technique to make use of the dataset. We also developed a method for incorporating sentence division, sentence merge, and sentence compression as components for single-document text summarization. We also developed a method for generating personalized snippets. We also developed a method for creating inning summaries for baseball matches. In the field of neural network-based summarization, we developed a method for controlling the output length for sequence-to-sequence summarization model, and a method for automatically creating a large-scale dataset for Japanese sentence compression.

Free Research Field

自然言語処理