Facilitating heuristic reasoning by utilizing knowledge graph and natural language processing
Haruna A. Noman K. Li Y. Makanda I.L.D. Zubair A. Hasand M.J. Alhassan A.B.
15 February 2026Elsevier B.V.
Knowledge-Based Systems
2026#334
Knowledge graphs (KG) have emerged as powerful tools for organizing and representing complex information, enabling machines to better understand and reason about data across various domains, including Additive Manufacturing (AM). However, in the context of AM, specifically in design for AM (DFAM), there is a lack of effective methods to utilize KGs for heuristic reasoning, which is vital for design workflows, resolving design-related obstacles, and boosting overall productivity. As a result, this study proposed a novel knowledge-based framework integrating KG and Natural Language Processing (NLP) techniques to facilitate heuristic reasoning in Fused Deposition (FDM) based DFAM. By leveraging text data from the AM-related “Thingiverse” platform, the approach constructs a KG that formalizes AM data into structured knowledge. The BERT model is utilized for entity and relation classification, while a modified Skip-gram neural model enhances the identification of semantic relationships within triples. Through the formulation of domain-specific design rules, the framework overcomes the challenges of heuristic reasoning efficiency and effectiveness in exploiting AMs capacity for product innovation and manufacturing expertise. To confirm the effectiveness of the proposed approach, a hands-on experiment and a real-world case study are both employed. The results demonstrate its superiority over traditional methods, with a remarkable accuracy of 97.62% and high F1 scores across various entity and relation categories. This indicates its effectiveness in knowledge classification and retrieval.
Additive manufacturing , Heuristic reasoning , Knowledge-based design, knowledge graph , Natural language processing , Skip-gram neural model
Text of the article Перейти на текст статьи
School of Aeronautics, Northwestern Polytechnical University, Xian, China
School of Civil Aviation, Northwestern Polytechnical University, Xian, China
Yangtze River Delta Research Institute of NPU, No. 27 Zigang Road, Science and Education New Town, Jiangsu Province, Taicang City, 215400, China
Aircraft Strength Research Institute of China, Xian, 710065, China
Applied AI, Dataxense, Guild Square, Aberdeen, Scotland, AB11 5RG, United Kingdom
Department of Robotics and Mechatronics, Nazarbayev University, Astana, 01000, Kazakhstan
School of Aeronautics
School of Civil Aviation
Yangtze River Delta Research Institute of NPU
Aircraft Strength Research Institute of China
Applied AI
Department of Robotics and Mechatronics
10 лет помогаем публиковать статьи Международный издатель
Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026