Knowledge Extraction via Decentralized Knowledge Graph Aggregation

摘要: In many industrial manufacturing processes, human operators play a central role when it comes to parameterizing the involved machinery and dealing with errors in process. However, large parts of acquired process knowledge are tacit, leading difficulties sharing between operators. Therefore, extraction is necessary but time cost intensive process, requiring both specially trained personnel experienced contrast, we propose that by gathering insights into what influenced operators' actual parameter choices, tacit can be extracted during production an example-based manner. This decentralized knowledge-decentralized regards who holds where was extracted—is then aggregated coherent graph. We showcase our methodology on real-world dataset domain fused deposition modeling (FDM), which generated providing their without additional assistance using extended machine interfaces. Furthermore, compare rules from graph against established FDM base showing viability approach even limited amounts data.