學習將自然語言敘述映射為知識圖譜表示形式以利知識庫之建立

Learning to Map Natural Language Statements into Knowledge Base Representations for Knowledge Base Construction

林勤和(Chin-Ho Lin)

指導教授 ： 陳信希

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2017年)

https://doi.org/10.6342/NTU201703086

摘要

透過資訊擷取系統以自動化方式建立知識庫，仍面臨許多挑戰。其中，對應知識在自然語言與知識庫中的字詞表示形式差異仍是一項問題。本篇論文以機器自動學習的方法，將知識三元組中的關係片語從自然語言形式的關係樣式，轉換為知識庫格式的關係謂語。我們在向量空間上訓練一個字詞表示模型，並藉此建立連結語意相等的關係樣式與關係謂語。相較於前人們的研究，我們所建立的關係片語對應表，不僅準確而且達到可觀的高覆蓋率。

關鍵字

知識庫建立 ； 關係片語 ； 關係片語映射 ； 字詞表示形式

並列摘要

Directly adding the knowledge triples obtained from open information extraction systems into a knowledge base is often impractical due to a vocabulary gap between natural language expressions and knowledge base representation. This thesis aims at learning to map relational phrases in triples from natural-language-like statement to knowledge base predicate format. We train a word representation model on a vector space and link each natural language relational pattern to semantically equivalent knowledge base predicate. Our mapping result shows not only high quality, but also promising coverage on relational phrases compared to previous researches.

並列關鍵字

Knowledge Base Construction ； Relational Phrases ； Relational Mapping ； Word Representation

參考文獻

Arnab Dutta, Christian Meilicke, and Heiner Stuckenschmidt. 2015. Enriching structured knowledge with open information. In WWW 2015. ACM, 267–277.

Kazuma Hashimoto and Yoshimasa Tsuruoka. 2016. Adaptive Joint Learning of Compositional and Non-Compositional Phrase Embeddings. In ACL 2016.

Johannes Kirschnick, Holmer Hemsen, and Volker Mark. 2016. JEDI: Joint Entity and Relation Detection using Type Inference. In ACL 2016, pages 61–66

Arvind Neelakantan, and Ming-Wei Chang. 2015. Inferring Missing Entity Type Instances for Knowledge Base Completion: New Dataset and Methods. In Proceedings of the 2015 Annual Conference of the North American Chapter of the ACL, pages 515–525

Jeffrey Pennington, Richard Socher, and Christopher D. Manning. 2014. GloVe: Global Vectors for Word Representation. In EMNLP 2014.