Abstract
We introduce a dynamical spatio-temporal model formalized as a recurrent neural network for modeling time series of spatial processes, i.e., series of observations sharing temporal and spatial dependencies. The model learns these dependencies through a structured latent dynamical component, while a decoder predicts the observations from the latent representations. We consider several variants of this model, corresponding to different prior hypothesis about the spatial relations between the series. The model is used for the tasks of forecasting and data imputation. It is evaluated and compared to state-of-the-art baselines, on a variety of forecasting and imputation problems representative of different application areas: epidemiology, geo-spatial statistics, and car traffic prediction. The experiments also show that this approach is able to learn relevant spatial relations without prior information.
Similar content being viewed by others
Notes
We assume that all the series have the same dimensionality and length. This is often the case for spatio-temporal problems otherwise this restriction can be easily removed.
In the experiments, we used the Nesterov’s Accelerated Gradient (NAG) method [36].
Code available at https://github.com/edouardelasalles/stnn.
We also performed tests with LSTM and obtained similar results as with GRU.
References
Bahadori MT, Yu QR, Liu Y (2014) Fast multivariate spatio-temporal analysis via low rank tensor learning. In: Advances in neural information processing systems, pp 3491–3499
Bańbura M, Modugno M (2014) Maximum likelihood estimation of factor models on datasets with arbitrary pattern of missing data. J Appl Econom 29(1):133–160
Bayer J, Osendorfer C (2014) Learning stochastic recurrent networks. arXiv preprint arXiv:1411.7610
Ben Taieb S, Hyndman R (2014) Boosting multi-step autoregressive forecasts. In: Proceedings of the 31st international conference on machine learning, pp 109–117
Bengio Y (2008) Neural net language models. Scholarpedia 3(1):3881
Ceci M, Corizzo R, Fumarola F, Malerba D, Rashkovska A (2017) Predictive modeling of pv energy production: How to set up the learning task for a better prediction? IEEE Trans Ind Inf 13(3):956–966
Che Z, Purushotham S, Cho K, Sontag D, Liu Y (2016) Recurrent neural networks for multivariate time series with missing values. arXiv preprint arXiv:1606.01865
Cho K, Van Merriënboer B, Gulcehre C, Bahdanau D, Bougares F, Schwenk H, Bengio Y (2014) Learning phrase representations using RNN encoder–decoder for statistical machine translation. arXiv preprint arXiv:1406.1078
Chung J, Kastner K, Dinh L, Goel K, Courville AC, Bengio Y (2015a) A recurrent latent variable model for sequential data. In: Advances in neural information processing systems, pp 2962–2970
Chung J, Kastner K, Dinh L, Goel K, Courville AC, Bengio Y (2015b) A recurrent latent variable model for sequential data. In: Cortes C, Lawrence ND, Lee DD, Sugiyama M, Garnett R (eds) Advances in neural information processing systems, vol 28. Curran Associates, Inc., Red Hook, pp 2980–2988
Connor JT, Martin RD, Atlas LE (1994) Recurrent neural networks and robust time series prediction. IEEE Trans Neural Netw 5(2):240–254
Cressie NAC, Wikle CK (2011) Statistics for spatio-temporal data, Wiley series in probability and statistics. Wiley, Hoboken
de Bezenac E, Pajot A, Gallinari P (2017) Deep learning for physical processes: incorporating prior scientific knowledge. In: Proceedings of the international conference on learning representations (ICLR)’
De Gooijer JG, Hyndman RJ (2006) 25 years of time series forecasting. Int J Forecast 22(3):443–473
Denton EL, Birodkar v (2017) Unsupervised learning of disentangled representations from video. In: Guyon I, Luxburg UV, Bengio S, Wallach H, Fergus R, Vishwanathan S, Garnett R (eds) Advances in Neural Information Processing Systems 30. Curran Associates, Inc., pp 4417–4426
Dornhege G, Blankertz B, Krauledat M, Losch F, Curio G, robert Muller K (2005) Optimizing spatio-temporal filters for improving brain-computer interfacing. In: in NIPS
Ganeshapillai G, Guttag J, Lo A (2013) Learning connections in financial time series. In: Proceedings of the 30th international conference on machine learning (ICML-13), pp 109–117
Graves A, Mohamed A-R, Hinton G (2013) Speech recognition with deep recurrent neural networks. In: IIIE ICASSP
Hochreiter S, Schmidhuber J (1997) Long short-term memory. Neural Comput 9(8):1735–1780
Honaker J, King G, Blackwell Matthew EA (2011) Amelia II: a program for missing data. J Stat Softw 45(7):1–47
Kalchbrenner N, Oord Avd, Simonyan K, Danihelka I, Vinyals O, Graves A, Kavukcuoglu K (2017) Unsupervised learning of video representations using lstms. In: Proceedings of the 34nd ICML-17
Kingma DP, Welling M (2013) Auto-encoding variational bayes. In: Proceedings of the 2nd international conference on learning representations (ICLR)
Koppula H, Saxena A (2013) Learning spatio-temporal structure from RGB-d videos for human activity detection and anticipation. In: Proceedings of ICML
Krishnan RG, Shalit U, Sontag D (2015) Deep kalman filters. arXiv preprint arXiv:1511.05121
Li Y, Tarlow D, Brockschmidt M, Zemel R (2015) Gated graph sequence neural networks. arXiv preprint arXiv:1511.05493
Mirowski P, LeCun Y (2009) Dynamic factor graphs for time series modeling. In: Buntine W, Grobelnik M, Mladenić D, Shawe-Taylor J (eds) Machine learning and knowledge discovery in databases. Springer, Berlin, Heidelberg, pp 128–143
Muller KR, Smola AJ, Ratsch G, Scholkopf B, Kohlmorgen J, Vapnik V (1999) Using support vector machines for time series prediction. Advances in kernel methods—support vector learning
Oord AV, Kalchbrenner N, Kavukcuoglu K (2016) Pixel recurrent neural networks, In: Balcan MF, Weinberger KQ (eds) Proceedings of The 33rd international conference on machine learning, vol. 48 of proceedings of machine learning research. PMLR, pp 1747–1756
Ren Y, Wu Y (2014) Convolutional deep belief networks for feature extraction of EEG signal. In: 2014 international joint conference on neural networks (IJCNN). IEEE, pp 2850–2853
Rue H, Held L (2005) Gaussian Markov random fields: theory and applications. CRC Press, Boca Raton
Shang J, Zheng Y, Tong W, Chang E, Yu Y (2014) Inferring gas consumption and pollution emission of vehicles throughout a city. In: Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, pp 1027–1036
Shi W, Zhu Y, Philip SY, Huang T, Wang C, Mao Y, Chen Y (2016) Temporal dynamic matrix factorization for missing data prediction in large scale coevolving time series. IEEE Access 4:6719–6732
Shi X, Chen Z, Wang H, Yeung D-Y, Wong W-k, Woo W-c (2015) Convolutional lstm network: a machine learning approach for precipitation nowcasting. In: Cortes C, Lawrence ND, Lee DD, Sugiyama M, Garnett R (eds) Advances in neural information processing systems, vol 28. Curran Associates, Inc., Red Hook, pp 802–810
Song Y, Liu M, Tang S, Mao X (2012) Time series matrix factorization prediction of internet traffic matrices. In: 2012 IEEE 37th conference on local computer networks (LCN). IEEE, pp 284–287
Srivastava N, Mansimov E, Salakhudinov R (2015) Unsupervised learning of video representations using lstms. In: Blei D, Bach F (eds) Proceedings of the 32nd ICML-15
Sutskever I, Martens J, Dahl G, Hinton G (2013) On the importance of initialization and momentum in deep learning. In: Proceedings of the 30th ICML
Sutskever I, Martens J, Hinton GE (2011) Generating text with recurrent neural networks. In: Proceedings of the 28th international conference on machine learning, ICML 2011
Wikle CK (2015) Modern perspectives on statistics for spatio-temporal data’. Wiley Interdiscip Rev Comput Stat 7(1):86–98
Wikle CK, Hooten MB (2010) A general science-based framework for dynamical spatio-temporal models’. Test 19(3):417–451
Yuan J, Zheng Y, Xie X, Sun G (2011) Driving with knowledge from the physical world. In: Proceedings of the 17th ACM SIGKDD. ACM, pp 316–324
Yuan J, Zheng Y, Zhang C, Xie W, Xie X, Sun G, Huang Y (2010) T-drive: driving directions based on taxi trajectories. In: Proceedings of the 18th SIGSPATIAL. ACM, pp 99–108
Acknowledgements
Locust Project ANR-15-CE23-0027-01, funded by Agence Nationale de la Recherche.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Delasalles, E., Ziat, A., Denoyer, L. et al. Spatio-temporal neural networks for space-time data modeling and relation discovery. Knowl Inf Syst 61, 1241–1267 (2019). https://doi.org/10.1007/s10115-018-1291-x
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10115-018-1291-x