Patent-KG: Patent Knowledge Graph Extraction for Engineering Design

Abstract This paper builds a patent-based knowledge graph, patent-KG, to represent the facts in patents for engineering design. The arising patent-KG approach proposes new unsupervised mechanism extract patent, by searching attention graph language models. extracted entities are compared with other benchmarks criteria of recall rate. result reaches highest 0.8 rate standard list mechanical related technical terms, which means coverage words.