GreenKGC: A Lightweight Knowledge Graph Completion Method

Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers, pages 10596–10613 July 9-14, 2023 ©2023 Association for Computational Linguistics GreenKGC: A Lightweight Knowledge Graph Completion Method Yun-Cheng Wang1, Xiou Ge1, Bin Wang2, C.-C. Jay Kuo1 1University of Southern California, Los Angeles, California, USA 2National University of Singapore, Singapore {yunchenw, xiouge, jckuo}@usc.edu, bwang28c@gmail.com Abstract Knowledge graph completion (KGC) aims to discover missing relationships between entities in knowledge graphs (KGs). Most prior KGC work focuses on learning embeddings for en- tities and relations through a simple scoring function. Yet, a higher-dimensional embedding space is usually required for a better reasoning capability, which leads to a larger model size and hinders applicability to real-world prob- lems (e.g., large-scale KGs or mobile/edge computing). A lightweight modularized KGC solution, called GreenKGC, is proposed in this work to address this issue. GreenKGC consists of three modules: representation learning, fea- ture pruning, and decision learning, to extract discriminant KG features and make accurate predictions on missing relationships using clas- sifiers and negative sampling. Experimental results demonstrate that, in low dimensions, GreenKGC can outperform SOTA methods in most datasets. In addition, low-dimensional GreenKGC can achieve competitive or even better performance against high-dimensional models with a much smaller model size. We make our code publicly available.1 1 Introduction Knowledge graphs (KGs) store human knowledge in a graph-structured format, where nodes and edges denote entities and relations, respectively. A (head entity, relation, tail entity) factual triple, de- noted by (h, r, t), is a basic component in KGs. In many knowledge-centric artificial intelligence (AI) applications, such as question answering (Huang et al., 2019; Saxena et al., 2020), information ex- traction (Hoffmann et al., 2011; Daiber et al., 2013), and recommendation (Wang et al., 2019; Xian et al., 2019), KG plays an important role as it provides explainable reasoning paths to predictions. How- ever, most KGs suffer from the incompleteness 1https://github.com/yunchengwang/Gree nKGC Figure 1: MRR versus the number of free parameters in KGE methods against FB15K-237 (left) and YAGO3-10 dataset (right). When a model has fewer parameters, its performance is poorer. Also, the larger dataset, YAGO3- 10, demands more parameters than the smaller dataset, FB15k-237, to achieve satisfactory results. problem; namely, a large number of factual triples are missing, leading to performance degradation in downstream applications. Thus, there is grow- ing interest in developing KG completion (KGC) methods to solve the incompleteness problem by inferring undiscovered factual triples based on ex- isting ones. Knowledge graph embedding (KGE) methods have been widely used to solve the in- completeness problem. Embeddings for entities and relations are stored as model parameters and updated by maximizing triple scores among ob- served triples while minimizing those among neg- ative triples. The number of free parameters in a KGE model is linear to the embedding dimen- sion and the number of entities and relations in KGs, i.e. O((|E| + |R|)d), where |E| is the num- ber of entities, |R| is the number of relations, and d is the embedding dimension. …

Content truncated to 500 words