Data augmentation and simultaneous prediction of optical properties of a porous core single mode fiber: CTGAN in the domain of optics
-
Afiquer Rahman
and
Md. Aslam Mollah
Abstract
Photonic crystal fiber (PCF) architectures have garnered significant interest due to their versatile applications across various fields. However, the complexity of PCF designs and the computational demands of full vectorial finite element method (FV-FEM) simulations pose challenges in fully realizing their potential. While prior research has explored artificial neural networks (ANNs) to accelerate simulation predictions, proposed approaches were often limited by sample size and generalizability. In this manuscript we introduce conditional generative adversarial networks (CTGAN) to augment real datasets, facilitating more effective ANN training. We evaluate CTGAN’s performance by comparing the quality of augmented data with real data and assess the predictive accuracy of ANNs trained on both augmented and real datasets. Our results demonstrate enhanced predictive accuracy, with higher R2 values for optical property predictions from structural parameters using the augmented dataset-trained ANN. Furthermore, the mean squared error (MSE) during ANN training decreased significantly (from 0.0051 to 0.0011), requiring fewer convergence epochs of only 88 compared to 114 for the real dataset. The proposed approach enables faster optical property predictions while reducing the required dataset generation simulations by up to 27.9 %.
Acknowledgments
We would like to express our gratitude to Department of Electronics & Telecommunication Engineering, Rajshahi University of Engineering & Technology, Rajshahi, Bangladesh.
-
Research ethics: We confirm that ethical approval does not apply to this article, given the absence of involvement with human or animal subjects.
-
Informed consent: Not applicable.
-
Author contributions: Conceptualization, A.R.; methodology, A.R., software, A.R., M.A.M.; validation, M.A.M.; formal analysis, M.A.M.; investigation, M.A.M.; resources, A.R.; data curation, A.R.; writing – original draft preparation, A.R.; writing – review and editing, A.R., M.A.M.; project administration, M.A.M; All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
-
Use of Large Language Models, AI and Machine Learning Tools: None declared.
-
Conflict of interest: The authors state no conflict of interest.
-
Research funding: This research is not funded from anywhere. There has no significant financial support for this work that could have influenced in its outcome.
-
Data availability: Data underlying the results presented in this paper is not publicly available but may be obtained from the authors upon reasonable request.
References
1. Mohamed Nizar, S, Rafi Ahamed, S, Priyanka, E, Jayasri, R, Kesavaraman, B. Comparison of different photonic crystal fiber structure: a review. J Phys Conf 2021;1717:012048. https://doi.org/10.1088/1742-6596/1717/1/012048.Search in Google Scholar
2. Leal Junior, AG, Frizera, A, Pontes, MJ. Analytical model for a polymer optical fiber under dynamic bending. Opt Laser Technol 2017;93:92–8. https://doi.org/10.1016/j.optlastec.2017.02.009.Search in Google Scholar
3. Dhatt, G, Lefrançois, E, Touzot, G. Finite element method. Hoboken, NJ: John Wiley & Sons; 2012.10.1002/9781118569764Search in Google Scholar
4. Li, H, Chen, H, Li, Y, Chen, Q, Li, S, Ma, M. Deep neural network for microstructured polymer fiber modeling. J Phys Appl Phys 2023;56:075104. https://doi.org/10.1088/1361-6463/acb369.Search in Google Scholar
5. Jabin, MA, Fok, MP. Prediction of 12 photonic crystal fiber optical properties using MLP in deep learning. IEEE Photon Technol Lett 2022;34:391–4. https://doi.org/10.1109/lpt.2022.3157266.Search in Google Scholar
6. Kumar, H, Jain, T, Sharma, M, Kishor, K. Neural network approach for faster optical properties predictions for different PCF designs. J Phys Conf 2021;2070:012001. https://doi.org/10.1088/1742-6596/2070/1/012001.Search in Google Scholar
7. Meng, F, Zhao, X, Ding, J, Niu, Y, Zhang, X, Smietana, M, et al.. Use of machine learning to efficiently predict the confinement loss in anti-resonant hollow-core fiber. Opt Lett 2021;46:1454–7. https://doi.org/10.1364/ol.422511.Search in Google Scholar PubMed
8. Kalyoncu, C, Yasli, A, Ademgil, H. Machine learning methods for estimating bent photonic crystal fiber based SPR sensor properties. Heliyon 2022;8. https://doi.org/10.1016/j.heliyon.2022.e11582.Search in Google Scholar PubMed PubMed Central
9. Shorten, C, Khoshgoftaar, TM. A survey on image data augmentation for deep learning. J Big Data 2019;6:1–48. https://doi.org/10.1186/s40537-019-0197-0.Search in Google Scholar
10. Ding, K, Xu, Z, Tong, H, Liu, H. Data augmentation for deep graph learning: a survey. ACM SIGKDD Explor Newsl 2022;24:61–77. https://doi.org/10.1145/3575637.3575646.Search in Google Scholar
11. Xu, L, Skoularidou, M, Cuesta-Infante, A, Veeramachaneni, K. Modeling tabular data using conditional GAN. Adv Neural Inf Process Syst 2019;32.Search in Google Scholar
12. Habibi, O, Chemmakha, M, Lazaar, M. Imbalanced tabular data modelization using CTGAN and machine learning to improve IoT Botnet attacks detection. Eng Appl Artif Intell 2023;118:105669. https://doi.org/10.1016/j.engappai.2022.105669.Search in Google Scholar
13. Ma, H, Geng, M, Wang, F, Zheng, W, Ai, Y, Zhang, W. Data augmentation of a corrosion dataset for defect growth prediction of pipelines using conditional tabular generative adversarial networks. Materials 2024;17:1142. https://doi.org/10.3390/ma17051142.Search in Google Scholar PubMed PubMed Central
14. Lee, JS, Lee, O. CTGAN vs TGAN? Which one is more suitable for generating synthetic EEG data. J Theor Appl Inf Technol 2021;99:2359–72.Search in Google Scholar
15. Islam, R, Hasanuzzaman, G, Habib, MS, Rana, S, Khan, M. Low-loss rotated porous core hexagonal single-mode fiber in THz regime. Opt Fiber Technol 2015;24:38–43. https://doi.org/10.1016/j.yofte.2015.04.006.Search in Google Scholar
16. Trisha, IJ, Patwary, AK, Sayem, MA, Rahman, A, Hossain, MR, Mollah, MA. Numerical analysis of a single channel exposed core elliptical shaped PCF based highly sensitive SPR sensor for wide RI sensing. Opt Express 2024;32:25472–87. https://doi.org/10.1364/oe.530030.Search in Google Scholar PubMed
17. Mortensen, NA. Effective area of photonic crystal fibers. Opt Express 2002;10:341–8. https://doi.org/10.1364/oe.10.000341.Search in Google Scholar PubMed
18. Kotsiantis, SB, Kanellopoulos, D, Pintelas, PE. Data preprocessing for supervised learning. Int J Comput Sci 2006;1:111–17. https://doi.org/10.4304/jcp.1.4.30-37.Search in Google Scholar
19. Gomez, S. Demystifying the CTGAN loss function: synthetic data modeling ⋅ SDV-Dev SDV ⋅ discussion #980, GitHub, sdv-team; 2022. Available from: https://github.com/sdv-dev/SDV/discussions/980.Search in Google Scholar
20. Abdi, H, Williams, LJ. Principal component analysis. Wiley Interdiscip Rev: Comput Stat 2010;2:433–59. https://doi.org/10.1002/wics.101.Search in Google Scholar
21. Dongare, AD, Kharde, RR, Kachare, AD. Introduction to artificial neural network. Int J Eng Innovative Technol (IJEIT) 2012;2:189–94.Search in Google Scholar
22. Bottou, L. Stochastic gradient descent tricks. In: Neural networks: tricks of the trade, 2nd ed. Berlin: Springer; 2012:421–36 pp.10.1007/978-3-642-35289-8_25Search in Google Scholar
23. Zhang, Q, Zhang, Y, Shao, Y, Liu, M, Li, J, Yuan, J, et al.. Boosting adversarial attacks with nadam optimizer. Electronics 2023;12:1464. https://doi.org/10.3390/electronics12061464.Search in Google Scholar
24. Singh, SK, Jain, SK, Bárdossy, A. Training of artificial neural networks using information-rich data. Hydrology 2014;1:40–62. https://doi.org/10.3390/hydrology1010040.Search in Google Scholar
25. Rahman, A, Mollah, MA. Reverse engineering and analysis of microstructure polymer fiber via artificial neural networks: simplifying the design approach. J Opt Commun 2024;45:2835–47. https://doi.org/10.1515/joc-2023-0361.Search in Google Scholar
Supplementary Material
This article contains supplementary material (https://doi.org/10.1515/joc-2024-0281).
© 2025 Walter de Gruyter GmbH, Berlin/Boston
