A Deeper Look into ‘Deep Learning of Aftershock Patterns Following Large Earthquakes’: Illustrating First Principles in Neural Network Physical Interpretability

  • Arnaud MignanEmail author
  • Marco Broccardo
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11506)


In the last years, deep learning has solved seemingly intractable problems, boosting the hope to find (approximate) solutions to problems that now are considered unsolvable. Earthquake prediction - a recognized moonshot challenge - is obviously worthwhile exploring with deep learning. Although encouraging results have been obtained recently, deep neural networks (DNN) may sometimes create the illusion that patterns hidden in data are complex when this is not necessarily the case. We investigate the results of De Vries et al. [Nature, vol. 560, 2018] who defined a DNN of 6 hidden layers with 50 nodes each, and with an input layer of 12 stress features, to predict aftershock patterns in space. The performance of their DNN was assessed using ROC with AUC = 0.85 obtained. We first show that a simple artificial neural network (ANN) of 1 hidden layer yields a similar performance, suggesting that aftershock patterns are not necessarily highly abstract objects. Following first principle guidance, we then bypass the elastic stress change tensor computation, making profit of the tensorial nature of neural networks. AUC = 0.85 is again reached with an ANN, now with only two geometric and kinematic features. Not only seems deep learning to be “excessive” in the present case, the simpler ANN streamlines the process of aftershock forecasting, limits model bias, and provides better insights into aftershock physics and possible model improvement. Complexification is a controversial trend in all of Science and first principles should be applied wherever possible to gain physical interpretations of neural networks.


Aftershock modelling Pattern recognition Applied deep learning 


  1. 1.
    LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521, 436–444 (2015)CrossRefGoogle Scholar
  2. 2.
    Jordan, M.I., Mitchell, T.M.: Machine learning: trends, perspectives, and prospects. Science 349(6245), 255–260 (2015)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Carleo, G., Troyer, M.: Solving the quantum many-body problem with artificial neural networks. Science 355, 602–606 (2017)MathSciNetCrossRefGoogle Scholar
  4. 4.
    Han, J., Jentzen, A., Weinan, E.: Solving high-dimensional partial differential equations using deep learning. PNAS 115(34), 8505–8510 (2018)MathSciNetCrossRefGoogle Scholar
  5. 5.
    Pathak, J., Hunt, B., Girvan, M., Lu, Z., Ott, E.: Model-free prediction of large spatiotemporally chaotic systems from data: a reservoir computing approach. Phys. Rev. Lett. 120, 024102 (2018)CrossRefGoogle Scholar
  6. 6.
    Kong, Q., Trugman, D.T., Ross, Z.E., Bianco, M.J., Meade, B.J., Gerstoft, P.: Machine learning in seismology: turning data into insights. Seismol. Res. Lett. 90(1), 3–14 (2019)CrossRefGoogle Scholar
  7. 7.
    Panakkat, A., Adeli, H.: Recurrent neural network for approximate earthquake time and location prediction using multiple seismicity indicators. Comput.-Aided Civ. Infrastruct. Eng. 24, 280–292 (2009)CrossRefGoogle Scholar
  8. 8.
    Geller, R.J., Jackson, D.D., Kagan, Y.Y., Mulargia, F.: Earthquakes cannot be predicted. Science 275(5306), 1616–1617 (1997)CrossRefGoogle Scholar
  9. 9.
    Brodi, B.: A neural-network model for earthquake occurrence. J. Geodyn. 32, 289–310 (2001)CrossRefGoogle Scholar
  10. 10.
    Moustra, M., Avraamides, M., Christodoulou, C.: Artificial neural networks for earthquake prediction using time series magnitude data or seismic electric signals. Expert Syst. Appl. 38, 15032–15039 (2011)CrossRefGoogle Scholar
  11. 11.
    DeVries, P.M.R., Viégas, F., Wattenberg, M., Meade, B.J.: Deep learning of aftershock patterns following large earthquakes. Nature 560, 632–634 (2018)CrossRefGoogle Scholar
  12. 12.
    Vere-Jones, D., Ben-Zion, Y., Zuniga, R.: Statistical seismology. Pure Appl. Geophys. 162, 1023–1026 (2005)CrossRefGoogle Scholar
  13. 13.
    Mignan, A.: Retrospective on the Accelerating Seismic Release (ASR) hypothesis: controversy and new horizons. Tectonophysics 505, 1–16 (2011)CrossRefGoogle Scholar
  14. 14.
    Sornette, D.: Critical Phenomena in Natural Sciences, Chaos, Fractals, Selforganization and Disorder: Concepts and Tools. Springer, New York (2009). Scholar
  15. 15.
    Mignan, A.: Seismicity precursors to large earthquakes unified in a stress accumulation framework. Geophys. Res. Lett. 39, L21308 (2012)CrossRefGoogle Scholar
  16. 16.
    Mignan, A.: Static behaviour of induced seismicity. Nonlin. Process. Geophys. 23, 107–113 (2016)CrossRefGoogle Scholar
  17. 17.
    Mignan, A.: Utsu aftershock productivity law explained from geometric operations on the permanent static stress field of mainshocks. Nonlin. Process. Geophys. 25, 241–250 (2018)CrossRefGoogle Scholar
  18. 18.
    Tiampo, K.F., Shcherbakov, R.: Seismicity-based earthquake forecasting techniques: ten years of progress. Tectonophysics 522–523, 89–121 (2012)CrossRefGoogle Scholar
  19. 19.
    Mignan, A.: Modeling aftershocks as a stretched exponential relaxation. Geophys. Res. Lett. 42, 9726–9732 (2015)CrossRefGoogle Scholar
  20. 20.
    Richards-Dinger, K., Stein, R.S., Toda, S.: Decay of aftershock density with distance does not indicate triggering by dynamic stress. Nature 467, 583–586 (2010)CrossRefGoogle Scholar
  21. 21.
    Hainzl, S., Brietzke, G.B., Zöller, G.: Quantitative earthquake forecasts resulting from static stress triggering. J. Geophys. Res. 115, B11311 (2010)CrossRefGoogle Scholar
  22. 22.
    Båth, M.: Lateral inhomogeneities of the upper mantle. Tectonophysics 2(6), 483–514 (1965)CrossRefGoogle Scholar
  23. 23.
    Gerstenberger, M.C., Wiemer, S., Jones, L.M., Reasenberg, P.A.: Real-time forecasts of tomorrow’s earthquakes in California. Nature 435, 328–331 (2005)CrossRefGoogle Scholar
  24. 24.
    Lakkos, S., Hadjiprocopis, A., Compley, R., Smith, P.: A neural network scheme for earthquake prediction based on the seismic electric signals. In: Proceedings of the IEEE Conference on Neural Networks and Signal Processing, pp. 681–689. IEEE, Ermioni (1994)Google Scholar
  25. 25.
    Alves, E.I.: Notice on the predictability of earthquake occurrences. Memórias e Notícias 117, 51–61 (1994)Google Scholar
  26. 26.
    Liu, Y., Wang, Y., Li, Y., Zhang, B., Wu, G.: Earthquake prediction by RBF neural network ensemble. In: Yin, F.-L., Wang, J., Guo, C. (eds.) ISNN 2004. LNCS, vol. 3174, pp. 962–969. Springer, Heidelberg (2004). Scholar
  27. 27.
    Alves, E.I.: Earthquake forecasting using neural networks: results and future work. Nonlin. Dyn. 44, 341–349 (2006)CrossRefGoogle Scholar
  28. 28.
    Panakkat, A., Adeli, H.: Neural network models for earthquake magnitude prediction using multiple seismicity indicators. Int. J. Neural Syst. 17(1), 13–33 (2007)CrossRefGoogle Scholar
  29. 29.
    Martínez-Álvarez, F., Reyes, J., Morales-Esteban, A., Rubio-Escudero, C.: Determining the best set of seismicity indicators to predict earthquakes. Two case studies Chile and the Iberian Peninsula. Knowl.-Based Syst. 50, 198–210 (2013)CrossRefGoogle Scholar
  30. 30.
    Asencio-Cortés, G., Martínez-Álvarez, F., Morales-Esteban, A., Reyes, J.: A sensitivity study of seismicity indicators in supervised learning to improve earthquake prediction. Knowl.-Based Syst. 101, 15–30 (2016)CrossRefGoogle Scholar
  31. 31.
    Madahizadeh, R., Allamehzadeh, M.: Prediction of aftershocks distribution using artificial neural networks and its application on the May 12, 2008 Sichuan earthquake. JSEE 11(3), 111–120 (2009)Google Scholar
  32. 32.
    Rouet-Leduc, B., Hulbert, C., Lubbers, N., Barros, K., Humphreys, C.J., Johnson, P.A.: Machine learning predicts laboratory earthquakes. Geophys. Res. Lett. 44, 9276–9282 (2017)CrossRefGoogle Scholar
  33. 33.
    Leach, R., Dowla, F.: Earthquake early warning system using real-time signal processing. In: Proceedings of the 1996 IEEE Signal Processing Society Workshop, pp. 463–472. IEEE, Kyoto (1996)Google Scholar
  34. 34.
    Kong, Q., Allen, R.M., Schreier, L., Kwon, Y.-W.: MyShake: a smartphone seismic network for earthquake early warning and beyond. Sci. Adv. 2, e1501055 (2016)CrossRefGoogle Scholar
  35. 35.
    Perol, T., Gharbi, M., Denolle, M.: Convolutional neural network for earthquake detection and location. Sci. Adv. 4, e1700578 (2018)CrossRefGoogle Scholar
  36. 36.
    Ross, Z.E., Meier, M.-A., Hauksson, E.: P wave arrival picking and first-motion polarity determination with deep learning. J. Geophys. Res. Solid Earth 123, 5120–5129 (2018)CrossRefGoogle Scholar
  37. 37.
    Ross, Z.E., Yue, Y., Meier, M.-A., Hauksson, E.: Phaselink: a deep learning approach to seismic phase association. J. Geophys. Res. Solid Earth (2019). Scholar
  38. 38.
    Bouchon, M., Karabulut, H., Aktar, M., Özalaybey, S., Schmittbuhl, J., Bouin, M.P.: Extended nucleation of the 1999 Mw 7.6 Izmit earthquake. Science 331(6019), 877–880 (2011)CrossRefGoogle Scholar
  39. 39.
    Mignan, A.: The debate on the prognostic value of earthquake foreshocks: a meta-analysis. Sci. Rep. 4, 4099 (2014)CrossRefGoogle Scholar
  40. 40.
    Mignan, A.: Asymmetric Laplace mixture modelling of incomplete power-law distributions: application to ‘seismicity vision’. In: Arai, K., Kapoor, S. (eds.) CVC 2019. AISC, vol. 944, pp. 30–43. Springer, Cham (2019). Scholar
  41. 41.
    Valentine, A.P., Trampert, J.: Data space reduction, quality assessment and searching of seismograms: autoencoder networks for waveform data. Geophys. J. Int. 189, 1183–1202 (2012)CrossRefGoogle Scholar
  42. 42.
    Li, Z., Meier, M.-A., Hauksson, E., Zhan, Z., Andrews, J.: Machine learning seismic wave discrimination: application to earthquake early warning. Geophys. Res. Lett. 45, 4773–4779 (2018)CrossRefGoogle Scholar
  43. 43.
    International Seismological Center. Accessed 29 Jan 2019
  44. 44.
    Finite-Source Rupture Model Database. Accessed 29 Jan 2019
  45. 45.
    Okada, Y.: Surface deformation due to shear and tensile faults in a half-space. Bull. Seismol. Soc. Am. 75(4), 1135–1154 (1985)Google Scholar
  46. 46.
    King, G.C.P.: Fault interaction, earthquake stress changes, and the evolution of seismicity. Treatise Geophys. 4, 225–255 (2007)CrossRefGoogle Scholar
  47. 47.
    Nature News: Artificial intelligence nails predictions of earthquake aftershocks. Accessed 29 Jan 2019
  48. 48.
    The New York Times: A.I. is Helping Scientists Predict When and Where the Next Big Earthquake Will Be. Accessed 29 Jan 2019
  49. 49.
    Futurism: Google’s AI can help predict where earthquake aftershocks are most likely. Accessed 29 Jan 2019
  50. 50.
    The Verge: Google and Harvard team up to use deep learning to predict earthquake aftershocks. Accessed 29 Jan 2019
  51. 51.
    Meade, B.J., DeVries, P.M.R., Faller, J., Viegas, F., Wattenberg, M.: What is better than coulomb failure stress? A ranking of scalar static stress triggering mechanisms from 105 mainshock-aftershock pairs. Geophys. Res. Lett. 44, 11409–11416 (2017)CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Swiss Federal Institute of TechnologyZurichSwitzerland

Personalised recommendations