Presentation 2023-05-26
グラフ文書における重要ノード抽出に関する検討
Yuto Maehashi, Toshiki Murata, Takahiro Yamasaki,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A graph document is a representation of the logical structure of a document in terms of a graph. A graph document has nodes that represent sentences and edges that represent their discourse relations. Because the discourse relations between sentences are clearly indicated in a graph document, the contents of the document can be grasped more accurately. In addition, it has the advantage that the logical structure of the document can be accurately confirmed when the document is created, making it easier to create documents that are easier to understand. On the other hand, as the number of nodes in a graph document increases, the readability of the document decreases and the main points of the document become difficult to understand. To solve this problem, this study focuses on the extraction of important nodes in graph documents. We define important node extraction as an extractive summarization task and propose a method for graph document summarization. We evaluated the proposed method on a Japanese summarization dataset and found that We confirmed that the proposed method is capable of summarizing using only information about the graph structure and labels assigned to edges.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Graph Document / Document Summarization / Extractive Summarization
Paper # LOIS2023-2
Date of Issue 2023-05-19 (LOIS)

Conference Information
Committee LOIS / IPSJ-SPT / IPSJ-CN
Conference Date 2023/5/26(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Hiroyuki Toda(NTT)
Vice Chair Manabu Motegi(Takushoku Univ.)
Secretary Manabu Motegi(Nagasaki Univ.) / (NTT)
Assistant Shuhei Yamamoto(NTT)

Paper Information
Registration To Technical Committee on Life Intelligence and Office Information Systems / Special Interest Group on Security Psychology and Trust / Collaboration and Network Services
Language JPN-ONLY
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English)
Sub Title (in English)
Keyword(1) Graph Document
Keyword(2) Document Summarization
Keyword(3) Extractive Summarization
1st Author's Name Yuto Maehashi
1st Author's Affiliation Oki Electric Industry Co., Ltd.(OKI)
2nd Author's Name Toshiki Murata
2nd Author's Affiliation Oki Electric Industry Co., Ltd.(OKI)
3rd Author's Name Takahiro Yamasaki
3rd Author's Affiliation Oki Electric Industry Co., Ltd.(OKI)
Date 2023-05-26
Paper # LOIS2023-2
Volume (vol) vol.123
Number (no) LOIS-55
Page pp.pp.9-12(LOIS),
#Pages 4
Date of Issue 2023-05-19 (LOIS)