Towards Time-Aware Knowledge Graph Completion

Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers, pages 1715–1724, Osaka, Japan, December 11-17 2016. Towards Time-Aware Knowledge Graph Completion Tingsong Jiang, Tianyu Liu, Tao Ge, Lei Sha, Baobao Chang, Sujian Li and Zhifang Sui Key Laboratory of Computational Linguistics, Ministry of Education School of Electronics Engineering and Computer Science, Peking University Collaborative Innovation Center for Language Ability, Xuzhou 221009 China {tingsong,tianyu0421,taoge,shalei,chbb,lisujian,szf}@pku.edu.cn Abstract Knowledge graph (KG) completion adds new facts to a KG by making inferences from existing facts. Most existing methods ignore the time information and only learn from time-unknown fact triples. In dynamic environments that evolve over time, it is important and challenging for knowledge graph completion models to take into account the temporal aspects of facts. In this paper, we present a novel time-aware knowledge graph completion model that is able to predict links in a KG using both the existing facts and the temporal information of the facts. To incorporate the happening time of facts, we propose a time-aware KG embedding model using temporal order information among facts. To incorporate the valid time of facts, we propose a joint time-aware inference model based on Integer Linear Programming (ILP) using temporal consistency information as constraints. We further integrate two models to make full use of global temporal information. We empirically evaluate our models on time-aware KG completion task. Experimental results show that our time-aware models achieve the state-of-the-art on temporal facts consistently. 1 Introduction Knowledge graphs (KGs) such as Freebase (Bollacker et al., 2008) and YAGO (Fabian et al., 2007) are extremely useful resources for many NLP related applications such as relation extraction and question answering, etc. Although KGs are large in size, they are far from complete (West et al., 2014). Knowl- edge graph completion, i.e., automatically inferring missing facts between entities in a knowledge graph, has thus become an increasingly important task. Recently a promising approach called KG embedding aims to embed the components (entities and relations) of a KG into a continuous vector space while preserving the inherent structure of a knowledge graph (Nickel et al., 2011; Bordes et al., 2011). This kind of approach has shown good effectiveness and scalability for KG completion. However, most existing KG embedding models ignore the temporal information of facts. In the real world, many facts are not static but highly ephemeral. For example, (Steve Jobs, diedIn, California) happened on 2011-10-05; (Ronaldo, playsFor, A.C.Milan) is true only during 2007-2008. Intuitively, temporal aspects of facts should play an important role when we perform KG completion. In this paper, we focus on time-aware KG completion. Specially, we incorporate two kinds of temporal information for KG completion: (a) temporal order information and (b) temporal consistency information. By temporal order information, we mean that many facts have temporal dependencies on others according to the time that they happened. For example, the facts involving a person P may follow the following timeline: (P, wasBornIn, ) → (P, graduateFrom, ) → (P, workAt, ) → (P, diedIn, ). Given …

Content truncated to 500 words