Coupling Corpus-Driven Knowledge Distillation with Microstructural Feature Learning for Predictive Material Modeling
DOI:
https://doi.org/10.5281/zenodo.20454747Keywords:
Automated Knowledge Extraction, Materials Science Knowledge Modeling, Microstructure–Property Relationships, Performance Prediction Models, Natural Language Processing Techniques, Classical Machine Learning Methods, Hybrid NLP–ML Framework, Pumpable Concrete Analysis, Knowledge Extraction Pipeline, Scientific Text Mining, SQLite Knowledge Database, Microstructure Feature Representation, Predictive Modeling In Materials, Relative Humidity Effects, Surface Resistivity Modeling, Electrical Conductivity Analysis, Binding Gel Volume Fraction, Concrete Cure Assessment, Data Benchmarking With Literature, Knowledge-Driven Materials EngineeringAbstract
Knowledge is vital in many domains, especially in study areas like materials science and engineering associated with the networks of interconnected factors and processes that determine a material's function. Such knowledge directly affects the ability to make informed forecasts. Materials selection and performance prediction models that are based on microstructure–property relationships require considerable knowledge about these dependencies. and the factors that govern them. Knowledge extraction, however, is both difficult and time-consuming. Automated knowledge extraction from natural language text can facilitate the development of such models.
A novel hybrid framework combines Natural Language Processing (NLP) techniques with classical Machine Learning (ML) methods to automate knowledge extraction from pumpable concrete publications. The Automated Knowledge Extraction (AKE) system encompasses two sub-processes, NLP and Microstructure Representation for Predictive Modeling. The first sub-process, Knowledge Extraction Pipeline, extracts knowledge from Relative Humidity–Surface Resistivity–Electrical Conductivity Relations for Accelerated Assessment of Concrete Cure paper and stores the extracted knowledge in an SQLite database. The second sub-process generates descriptive microstructure features suitable for predictive modeling with ML models. Predictive models calculated data have been benchmarked against data from the literature. The volume fraction of binding gel is the main influential factor that governs the Relationship between Relative Humidity and Surface Resistivity and the Relationship between Relative Humidity and Electrical Conductivity. The hybrid framework therefore enables automated knowledge extraction from natural language pumpable concrete publications and provides descriptive microstructure features that can be used for automated performance prediction with ML algorithms and classifiers.
References
[1] Bock, F. E., Aydin, R. C., Cyron, C. J., Huber, N., Kalidindi, S. R., & Klusemann, B. (2019). A review of the application of machine learning and data mining approaches in continuum materials mechanics. Frontiers in Materials, 6, 110.
[2] Butler, K. T., Davies, D. W., Cartwright, H., Isayev, O., & Walsh, A. (2018). Machine learning for molecular and materials science. Nature, 559(7715), 547–555.
[3] Chen, C., Zuo, Y., Ye, W., Li, X., Deng, Z., & Ong, S. P. (2020). A critical review of machine learning of energy materials. Advanced Energy Materials, 10(8), 1903242.
[4] Chen, X., Liu, Y., & Gu, G. X. (2023). Semantic-aware deep learning for structure–property relationships in advanced materials. Materials & Design, 230, 111977.
[5] Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT Press.
[6] Gu, G. X., Chen, C. T., Richmond, D. J., & Buehler, M. J. (2018). Bioinspired hierarchical composite design using machine learning. Materials Horizons, 5(5), 939–945.
[7] Haghighatlari, M., Li, J., Heidar-Zadeh, F., Liu, Y., Guan, X., & Head-Gordon, T. (2020). Learning to make chemical predictions. Chemistry of Materials, 32(17), 7375–7390.
[8] Jha, D., Ward, L., Paul, A., Liao, W. K., Choudhary, A., Wolverton, C., & Agrawal, A. (2018). ElemNet: Deep learning the chemistry of materials. Scientific Reports, 8, 17593.
[9] Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3(6), 422–440.
[10] Kim, E., Huang, K., Saunders, A., McCallum, A., & Olivetti, E. (2017). Materials synthesis insights from scientific literature. Chemistry of Materials, 29(21), 9436–9444.
[11] Kong, Z., Lu, J., Wang, Y., & Zhang, C. (2020). Machine learning-based fatigue life prediction of composite laminates. Composite Structures, 234, 111677.
[12] Liu, Y., Zhang, Z., Ling, C., & Chen, C. (2022). Knowledge-guided graph neural networks for materials property prediction. npj Computational Materials, 8(1), 125.
[13] Lu, L., Meng, X., Mao, Z., & Karniadakis, G. E. (2021). DeepXDE: A deep learning library for differential equations. SIAM Review, 63(1), 208–228.
[14] Mao, W., Zhang, J., & Yang, Z. (2021). Intelligent damage detection of composite structures using deep learning. Smart Materials and Structures, 30(7), 075018.
[15] Miotto, R., Wang, F., Wang, S., Jiang, X., & Dudley, J. T. (2018). Deep learning for healthcare: Review, opportunities and challenges. Briefings in Bioinformatics, 19(6), 1236–1246.
[16] Molnar, C. (2022). Interpretable machine learning (2nd ed.). Leanpub.
[17] Nguyen, T. T., Hoang, N. D., & Tran, X. T. (2022). Deep learning-based assessment of composite material degradation. Composite Structures, 290, 115558.
[18] Raccuglia, P., Elbert, K. C., Adler, P. D. F., Falk, C., Wenny, M. B., Mollo, A., Zeller, M., Friedler, S. A., Schrier, J., & Norquist, A. J. (2016). Machine-learning-assisted materials discovery using failed experiments. Nature, 533(7601), 73–76.
[19] Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). Why should I trust you? Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 1135–1144.
[20] Schmidt, J., Marques, M. R. G., Botti, S., & Marques, M. A. L. (2019). Machine learning in solid-state materials science. npj Computational Materials, 5, 83.
[21] Tao, F., Qi, Q., Liu, A., & Nee, A. Y. C. (2018). Digital twin-driven product design and manufacturing. International Journal of Advanced Manufacturing Technology, 94(9–12), 3563–3576.
[22] Ward, L., Agrawal, A., Choudhary, A., & Wolverton, C. (2016). A general-purpose machine learning framework for materials. npj Computational Materials, 2, 16028.
[23] Wen, C., Zhang, Y., & Hu, J. (2023). Transformer-based deep learning for composite materials. Composite Structures, 311, 116789.
[24] Wu, S., Chen, X., Zhang, Y., & Gu, G. X. (2022). Interpretable deep learning for composite material design. Materials & Design, 219, 110782.
[25] Xie, T., & Grossman, J. C. (2018). Crystal graph convolutional neural networks. Physical Review Letters, 120(14), 145301.
[26] Yang, Z., Yabansu, Y. C., Al-Bahrani, R., Liao, W. K., Choudhary, A. N., Kalidindi, S. R., & Agrawal, A. (2018). Deep learning for structure–property linkages in composites. Computational Materials Science, 151, 278–287.
[27] Zhang, Y., Ling, C., & Chen, C. (2021). A graph-based framework for material knowledge representation. Computational Materials Science, 196, 110525.
[28] Zhu, Y., Zabaras, N., Koutsourelakis, P. S., & Perdikaris, P. (2019). Physics-constrained deep learning. Journal of Computational Physics, 394, 56–81.
[29] Li, Y., Zhang, H., & Wang, L. (2022). Multimodal deep learning for intelligent materials characterization. Composite Structures, 296, 115896.
[30] Sun, C., Chen, Z., & Liu, J. (2023). Knowledge-enhanced neural networks for material performance prediction. Engineering Applications of Artificial Intelligence, 121, 105960.
[31] Zhao, Y., Wu, D., & Zhang, X. (2024). Semantic-guided deep learning for damage characterization of transportation materials. Composite Structures, 332, 117654.
[32] Liu, M., Frangopol, D. M., & Kim, S. (2019). Bridge deterioration modeling using Bayesian updating. Journal of Structural Engineering, 145(5), 04019023.
[33] Shen, D., Wu, G., & Suk, H. I. (2017). Deep learning in medical image analysis. Annual Review of Biomedical Engineering, 19, 221–248.
[34] Pan, S. J., & Yang, Q. (2010). A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 22(10), 1345–1359.
