Does William Shakespeare REALLY Write Hamlet? Knowledge Representation Learning With Confidence

• DOI: 10.1609/aaai.v32i1.11924
• Publication Date: 2022-06-24T21:33:48Z
• Authors: Ruobing Xie; Zhiyuan Liu; Fen Lin; Leyu Lin
• Keywords: Knowledge graph; Representation; Feature Learning

Knowledge graphs (KGs), which could provide essential relational information between entities, have been widely utilized in various knowledge-driven applications. Since the overall human knowledge is innumerable that still grows explosively and changes frequently, construction update inevitably involve automatic mechanisms with less supervision, usually bring plenty of noises conflicts to KGs. However, most conventional representation learning methods assume all triple facts existing KGs share same significance without any noises. To address this problem, we propose a novel confidence-aware framework (CKRL), detects possible while representations confidence simultaneously. Specifically, introduce translation-based for learning. make more flexible universal, only utilize internal structural KGs, three kinds confidences considering both local global information. In experiments, We evaluate our models on graph noise detection, completion classification. Experimental results demonstrate achieve significant consistent improvements tasks, confirms capability CKRL modeling KG detection