Name
Abstract | ||
|---|---|---|
Entity alignment is a basic and vital technique in knowledge graph (KG) integration. Over the years, research on entity alignment has resided on the assumption that KGs are static, which neglects the nature of growth of real-world KGs. As KGs grow, previous alignment results face the need to be revisited while new entity alignment waits to be discovered. In this paper, we propose and dive into a realistic yet unexplored setting, referred to as continual entity alignment. To avoid retraining an entire model on the whole KGs whenever new entities and triples come, we present a continual alignment method for this task. It reconstructs an entity's representation based on entity adjacency, enabling it to generate embeddings for new entities quickly and inductively using their existing neighbors. It selects and replays partial pre-aligned entity pairs to train only parts of KGs while extracting trustworthy alignment for knowledge augmentation. As growing KGs inevitably contain non-matchable entities, different from previous works, the proposed method employs bidirectional nearest neighbor matching to find new entity alignment and update old alignment. Furthermore, we also construct new datasets by simulating the growth of multilingual DBpedia. Extensive experiments demonstrate that our continual alignment method is more effective than baselines based on retraining or inductive learning. |
Year | DOI | Venue |
|---|---|---|
2022 | 10.1007/978-3-031-19433-7_12 | IEEE International Semantic Web Conference |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 7 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Yuxin Wang | 1 | 0 | 0.34 |
| Yuanning Cui | 2 | 0 | 0.34 |
| Wenqiang Liu | 3 | 0 | 0.68 |
| Zequn Sun | 4 | 52 | 9.00 |
| Yiqiao Jiang | 5 | 0 | 0.34 |
| Kexin Han | 6 | 0 | 0.68 |
| Yuzhong Qu | 7 | 726 | 62.49 |