Beyond expression: Comprehensive visualization of knowledge triplet facts

Open Webpage

Wei Liu, Yixue He, Chao Wang, Shaorong Xie, Weimin Li

Published: 01 Jan 2025, Last Modified: 20 May 2025Inf. Process. Manag. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Our RGE method achieves comprehensive visualization of triplet facts.•We thoroughly consider the grounding of non-visual relations in triplet facts.•We construct DB15kFact, a multi-modal knowledge graph with 86,722 triplets and 387,096 images.•DB15kFact enhances performance in link prediction and relation classification.