Research on Material Translation Matching Model for Regional Environmental Design Based on NLP and Cultural Gene Mapping

Jing Zhang1*
1 College of Arts, Shandong Jianzhu University, Jinan, Shandong, China
* Corresponding author: Jing Zhang. Email: outlook_8A9201AAC2554F63@outlook.com
International Scientific Technical and Economic Research 2026, Vol. 4, No. 3, pp. 60-89
DOI: 10.67541/istaer2626
Received: 30 June 2026; Revised: 18 July 2026; Accepted: 30 July 2026; Published: 11 August 2026
Abstract

The selection of materials for regional environment design has long been trapped in the cross-modal gap between cultural semantics and physical properties, and existing methods find it difficult to transform the tacit knowledge in the construction tradition into a computable matching basis. This paper proposes a material translation matching model based on natural language processing and cultural gene mapping. Through deep semantic extraction of multi-source heterogeneous corpora, a heterogeneous information network with "geography, history and technology" as hyperedges is constructed to realize the joint embedding representation of material nodes and cultural symbol nodes. Aiming at the fuzziness and dynamics of design intention, a multi-dimensional constraint tensor modeling method based on fuzzy membership function is designed, and Pareto front ranking and context-aware relaxation coefficient are introduced to realize online adaptive adjustment of elastic matching window. At the matching computing level, a dynamic graph attention aggregation network is proposed to capture semantic associations in heterogeneous structures through node-level, relationship-level and path-level attention mechanism, and integrate cross-modal contrast learning to align text constraint space and graph embedding space. Experiments are carried out on the standardized test set covering three regions, namely, water towns in the south of the Yangtze River, red brick in the south of Fujian, and forest pan in the west of Sichuan. The matching accuracy @1 of the complete model reaches 66.8%, which is 22.8 percentage points higher than the optimal baseline model. The lightweight variant achieves a single inference latency of 42.8 ms at the Jetson Xavier NX edge with an accuracy loss of only 6.0 percentage points. In the blind evaluation, 78.5% of the recommended results were judged as highly consistent by senior architects, and the relative deviation of physical and chemical indicators was within the engineering allowable ranges. Ablation experiments and generalization tests verify the effectiveness of each module and the ability of cross-regional migration of the model.

Keywords
Cultural gene mapping Heterogeneous information network Dynamic graph attention aggregation network Cross-modal contrast learning Regional environment design material matching
References
  1. Zheng, K. (2024). Regional cultural integration and innovation in environmental design: Exploring strategies to maintain and inherit regional cultural values in the context of globalization. Cultura: International Journal of Philosophy of Culture and Axiology, 21(3). DOI: 10.xxxx/cultura.2024.542
  2. Shou, T., Pan, Y., & Zhao, B. (2026). Vernacular ecological construction: a diagrammatic exploration on the composition modes of spatial-landscape in Huizhou courtyard dwellings. Architectural Engineering and Design Management, 22(2), 446-470. DOI: 10.1080/17452007.2025.2517136
  3. Paehler, L., & Matthiesen, S. (2024). Mapping the landscape of product models in embodiment design. Research in Engineering Design, 35(3), 289-310. DOI: 10.1007/s00163-024-00433-x
  4. Guérineau, J., Bricogne, M., Rivest, L., & Durupt, A. (2022). Organizing the fragmented landscape of multidisciplinary product development: a mapping of approaches, processes, methods and tools from the scientific literature. Research in Engineering Design, 33(3), 307-349. DOI: 10.1007/s00163-022-00389-w
  5. Wang, W. F., Lu, C. T., Ponsa i Campanyà, N., Chen, B. Y., & Chen, M. Y. (2025). AIdeation: Designing a human-AI collaborative ideation system for concept designers. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (pp. 1-28). DOI: 10.1145/3706598.3714148
  6. Ranjgar, B., Sadeghi-Niaraki, A., Shakeri, M., Rahimi, F., & Choi, S. M. (2024). Cultural heritage information retrieval: past, present, and future trends. IEEE Access, 12, 42992-43026. DOI: 10.1109/ACCESS.2024.3374769
  7. Hambarde, K. A., & Proenca, H. (2023). Information retrieval: recent advances and beyond. IEEE Access, 11, 76581-76604. DOI: 10.1109/ACCESS.2023.3295776
  8. Golub, K., & Szostak, R. (2025). Information retrieval of humanities resources: subject searching from a user perspective. Journal of Documentation, 81(7), 376-398. DOI: 10.1108/JD-05-2025-0129
  9. Ozdemir, A., Odaci, B., Tanatar Baruh, L., Varol, O., & Balcisoy, S. (2025). Enhancing cultural heritage archive analysis via automated entity extraction and graph-based representation learning. ACM Journal on Computing and Cultural Heritage, 18(4), 1-25. DOI: 10.1145/3746658
  10. Liang, Y., Xie, B., Tan, W., & Zhang, Q. (2025). Ontology-based construction of embroidery intangible cultural heritage knowledge graph: A case study of Qingyang sachets. PLoS One, 20(1), e0317447. DOI: 10.1371/journal.pone.0317447
  11. Zhao, T., & Na, R. (2026). Semantic Mapping and Cross-Model Data Integration in BIM: A Lightweight and Scalable Schedule-Level Workflow. Buildings, 16(7), 1347. DOI: 10.3390/buildings16071347
  12. Quek, H. Y., Hofmeister, M., Rihm, S. D., Yan, J., Lai, J., Brownbridge, G., ... & Kraft, M. (2024). Dynamic knowledge graph applications for augmented built environments through "The World Avatar". Journal of Building Engineering, 91, 109507. DOI: 10.1016/j.jobe.2024.109507
  13. Gao, C., Zheng, Y., Li, N., Li, Y., Qin, Y., Piao, J., ... & Li, Y. (2023). A survey of graph neural networks for recommender systems: Challenges, methods, and directions. ACM Transactions on Recommender Systems, 1(1), 1-51. DOI: 10.1145/3568022
  14. Wu, S., Sun, F., Zhang, W., Xie, X., & Cui, B. (2022). Graph neural networks in recommender systems: a survey. ACM Computing Surveys, 55(5), 1-37. DOI: 10.1145/3535101
  15. Anand, V., & Maurya, A. K. (2025). A survey on recommender systems using graph neural network. ACM Transactions on Information Systems, 43(1), 1-49. DOI: 10.1145/3694784
  16. Lyu, Z., Wu, Y., Lai, J., Yang, M., Li, C., & Zhou, W. (2022). Knowledge enhanced graph neural networks for explainable recommendation. IEEE Transactions on Knowledge and Data Engineering, 35(5), 4954-4968. DOI: 10.1109/TKDE.2022.3142260
  17. Hsu, S. F., Yang, C. H., Shidujaman, M., & Mikhailova, V. (2026). Can AI serve as a reading guide? Exploring how generative AI enhances information comprehension. International Journal of Human–Computer Interaction, 1-18. DOI: 10.1080/10447318.2026.2659306
  18. Bodon, H., Kumar, V., & Worsley, M. (2025). Design Principles for Authentically Embedding Computer Science in Sports. ACM Transactions on Computing Education, 25(2), 1-42. DOI: 10.1145/3722228
  19. Qiang, B., Xi, G., Wang, Y., Yang, X., & Wang, Y. (2023). Siamese Interaction and Fine‐Tuning Representation of Chinese Semantic Matching Algorithm Based on RoBERTa‐wwm‐ext. Mobile Information Systems, 2023(1), 8704278. DOI: 10.1155/2023/8704278
  20. Gao, F., Zhang, L., Wang, W., Zhang, B., Liu, W., Zhang, J., & Xie, L. (2024). Named entity recognition for equipment fault diagnosis based on RoBERTa-wwm-ext and deep learning integration. Electronics, 13(19), 3935. DOI: 10.3390/electronics13193935
  21. Dong, H., Kong, Y., Gao, W., & Liu, J. (2022). Named entity recognition for public interest litigation based on a deep contextualized pretraining approach. Scientific Programming, 2022(1), 7682373. DOI: 10.1155/2022/7682373
  22. Zhang, W., Ma, L., Xu, J., & Peng, Y. (2025). Spatial–temporal distribution characteristics and influencing factors of immovable cultural heritage in Hubei, China. Scientific Reports, 15(1), 24570. DOI: 10.1038/s41598-025-09764-8
  23. Gu, Y., Zhang, Y., Hou, Y., Yu, S., Li, G., Huang, H., & Su, D. (2025). The spatio-temporal characteristics and influencing factors of intangible cultural heritage in Jiang-Zhe-Hu region, China. Sustainability, 18(1), 35. DOI: 10.3390/su18010035
  24. Noori, A., Balafar, M. A., Bouyer, A., & Salmani, K. (2024). Review of heterogeneous graph embedding methods based on deep learning techniques and comparing their efficiency in node classification. Social Network Analysis and Mining, 14(1), 17. DOI: 10.1007/s13278-023-01178-6
  25. Han, J., Lu, X. Z., & Lin, J. R. (2025). Pretrained graph neural network for embedding semantic, spatial, and topological data in building information models. Computer‐Aided Civil and Infrastructure Engineering, 40(26), 4607-4631. DOI: 10.1111/mice.70073
  26. Chen, C., Li, Q., Chen, L., Liang, Y., & Huang, H. (2023). An improved GraphSAGE to detect power system anomaly based on time-neighbor feature. Energy Reports, 9, 930-937. DOI: 10.1016/j.egyr.2022.11.116
  27. Yang, Z., Yang, Z., Tian, W., Li, J., Sun, X., Zheng, G., ... & Zhong, W. (2026). FusionMAE, a self-supervised pretrained model to optimize and simplify diagnostic and control of fusion plasma. Communications Physics. DOI: 10.1038/s42005-026-02626-3
  28. Nguyen, D. H., Phan, N. D. M., & Pham, T. S. (2026). Parallel feature fusion with mixed normalization for efficient tool wear classification in CNC milling: HyMixNet. Engineering Research Express, 8(13), 135402. DOI: 10.1088/2631-8695/ae7d91