Application of data augmentation based on generative adversarial network in impedance inversion
Wang, P., Xu, H.Q., Peng, Z., Wang, Z. and Yang, M.Q., 2023. Application of data augmentation based on generative adversarial network in impedance inversion. Journal of Seismic Exploration, 32: 155-168. In recent years, various deep learning techniques have been widely used in the field of geophysics. As far as seismic impedance inversion is concerned, a nonlinear mapping model from seismic data to wave impedance can be established by training the depth inversion network, and then the impedance information can be predicted by the nonlinear model. However, the effectiveness of the current impedance inversion methods based on deep neural networks depends on the number of labels. The generalization ability of the model trained in the state of few labels is poor. Data augmentation can alleviate this situation by using the existing data. Therefore, the author proposes a method based on generative adversarial network (GAN) to augment the labels in the original data set, and uses geophysical forward modeling technology to forward seismic data to achieve the function of data augmentation. Unlike existing GAN, which generates samples directly from noise, the method augments the labeled data from the original dataset. The validity of this method is verified by model data and actual data. The method provides a data augmentation seismic inversion technique based on GAN for impedance inversion.
- Antoniou, A., Storkey, A. and Edwards, H., 2017. Data augmentation generativeadversarial networks. arXiv preprint arXiv:1711.04340.
- Ba, J.L., Kiros, J.R. and Hinton, G.E., 2016. Layer normalization. arXiv preprintarXiv: 1607.06450.
- Bai, S., Kolter, J.Z. and Koltun, V., 2018. An empirical evaluation of genericconvolutional and recurrent networks for sequence modeling. arXiv preprintarXiv: 1803.01271.
- Bayer, M., Kaufhold, M.A., Buchhold, B., Keller, M., Dallmeyer, J. and Reuter,
- C.. 2022. Data augmentation in natural language processing: a novel textgeneration approach for long and short text classifiers. Internat. J. Mach.Learning and Cybernetics, (Vol. No. ???): 1-16.
- Cai, A., Di, H., Li, Z., Maniar, H. and Abubakar, A., 2020. Wasserstein cycleconsistent generative adversarial network for improved seismic impedanceinversion: Example on 3D SEAM model. Expanded Abstr., 20th Ann. Internat.SEG Mtg., Houston: 1274-1278.
- Du, C. and Huang, L. , 2019. Sentiment analysis method based on piecewiseconvolutional neural network and generative adversarial network. Internat. J.Comput. Communicat. Contr., 14: 7-20.
- Feng, S.Y., Gangal, V., Wei, J., Chandar, S., Vosoughi, S., Mitamura, T. and
- Hovy, E., 2021. A survey of data augmentation approaches for NLP. arXivpreprint arXiv:2105.03075.
- Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S.,and Bengio, Y., 2014. Generative adversarial nets. Advan. Neural Inform.Process. Syst., 27: (PAGE Nos. ???)
- Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V. and Courville, A.C., 2017.
- Improved training of Wasserstein Gans. Advan. Neural Inform. Process. Syst.,30: (PAGE Nos. ???)
- Hinton, G.E., Srivastava, N., Krizhevsky, A., Sutskever, I. and Salakhutdinov, R.,
- Improving neural networks by preventing co-adaptation of featuredetectors. arXiv preprint arXiv:1207.0580.loffe, S. and Szegedy, C., 2015. Batch normalization: Accelerating deep networktraining by reducing internal covariate shift. Abstr. Internat. Conf. MachineLearn.: 448-456. PMLR.
- Ledig, C., Theis, L., Huszar, F., Caballero, J., Cunningham, A., Acosta, A. and
- Shi, W., 2017. Photo-realistic single image super-resolution using a generativeadversarial network. Proc. IEEE Conf. Comput. Vision Pattern Recognit.: 4681-4690.
- Mikolajezyk, A. and Grochowski, M., 2018. Data augmentation for improving deeplearning in image classification problem. IEEE Internat. Interdiscipl. Ph.D.workshop (IIPhDW): 117-122.
- Nie, D., Trullo, R., Lian, J., Petitjean, C., Ruan, S., Wang, Q. and Shen, D., 2017.
- Medical image synthesis with context-aware generative adversarial networks.
- Internat. Conf. Medic. Image Comput. Computer-assisted Intervent.: 417-425.Springer, New York.
- Pascual, S., Bonafonte, A. and Serra, J., 2017. SEGAN: Speech Enhancementgenerative adversarial network. arXiv preprint arXiv: 1703.09452.
- Razzak, M.I., Naz, S. and Zaib, A., 2018. Deep learning for medical imageprocessing: Overview, challenges and the future. Classificat. BioApps, 323-350.
- Salimans, T. and Kingma, D.P., 2016. Weight normalization: A simple reparamet-erization to accelerate training of deep neural networks. Advan. Neural Infor.Process. Syst., 29.
- Shorten, C. and Khoshgoftaar, T.M., 2019. A survey on image data augmentationfor deep learning. J. Big Data, 6: 1-48.
- Suh, S., Lee, H., Lukowicz, P. and Lee, Y.O., 2021. CEGAN: Classification
- Enhancement Generative Adversarial Networks for unraveling data imbalanceproblems. Neural Netw., 133: 69-86.
- Wang, X., Zhao, Y. and Pourpanah, F., 2020. Recent advances in deep learning.Internat. J. Mach. Learn. Cybernet., 11: 747-750.
- Wang, Y.Q., Wang, Q., Lu, W.K., Ge, Q. and Yan, X.F., 2022. Seismic impedanceinversion based on cycle-consistent generative adversarial network. Petrol. Sci.,19: 147-161.
- Yi, X.D., Wu, B.Y., Meng, D. and Cao, X.Y., 2021. Application of data augmen-tation and active learning to seismic wave impedance inversion. Oil Geophys.
- Prosp., 56: 707-715. DOI: 10.13810/).cnki.issn.1000-7210.2021.04.004.
- Young, T., Hazarika, D., Poria, S. and Cambria, E., 2018. Recent trends in deeplearning based natural language processing. IEEE Computat. Intellig. Magaz.,13:55-75.
