TeDS: Joint Learning of Diachronic and Synchronic Perspectives in Quaternion Space for Temporal Knowledge Graph Completion

Jiujiang Guo, Mankun Zhao, Wenbin Zhang, Tianyi Xu, Linying Xu, Yu Jian, Yu Mei, Yu Ruiguo

TL;DR: The paper simulates and summarizes dual perspectives of temporalized facts and designs distinct temporal perceptions based on quaternions to enhance TKGC.

Abstract: Existing research on temporal knowledge graph completion treats temporal information as supplementary, without simulating various features of facts from a temporal perspective. This work summarizes features of temporalized facts from both diachronic and synchronic perspectives: (1) Diachronicity. Facts often exhibit varying characteristics and trends across different temporal domains; (2) Synchronicity. In specific temporal contexts, various relations between entities influence each other, generating latent semantics. To track above issues, we design a quaternion-based model, TeDS, which divides timestamps into diachronic and synchronic timestamps to support dual temporal perception: (a) Two composite quaternions fusing time and relation information are generated by reorganizing synchronic timestamp and relation quaternions, and Hamilton operator achieves their interaction. (b) Each time point is sequentially mapped to an angle and converted to scalar component of a quaternion using trigonometric functions to build diachronic timestamps. We then rotate relation by using Hamilton operator between it and diachronic timestamp. In this way, TeDS achieves deep integration of relations and time while accommodating different perspectives. Empirically, TeDS significantly outperforms SOTA models on six benchmarks.

Lay Summary: (1) High-quality knowledge graphs (KGs) enable faster and more accurate extraction of key information, enhancing the efficiency of intelligent retrieval, reasoning, and decision-making. However, the quality of KGs often heavily depends on the methods of information acquisition, the scope of coverage, and the timeliness of information. These factors can lead to fluctuations in quality, which significantly affect the effectiveness and scalability of KGs in real-world applications. Therefore, it is essential to develop more refined and intelligent construction and completion mechanisms to support the sustainable development and efficient use of high-quality KGs. (2) This study abstracts the characteristic rules of factual knowledge in the physical world by drawing from real-world patterns: (a) **Diachronicity** Facts often exhibit different features and developmental trends across various time periods. (b) **Synchronicity** In specific temporal contexts, multiple relations between entities interact with each other, thereby generating latent semantics. Furthermore, we propose a quaternion-based model, TeDS, which divides timestamps into diachronic and synchronic timestamps to support dual temporal perception. (3) This will help simulate and understand knowledge graph structures from a real-world perspective, enhance the interpretability and expressiveness of knowledge representation, and further provide sustainable solutions for knowledge graph completion and construction, empowering practical applications.

Application-Driven Machine Learning: This submission is on Application-Driven Machine Learning.

Primary Area: General Machine Learning->Representation Learning

Keywords: Knowledge graph; Knowledge graph completion

Submission Number: 1610